Hand your coding agent a management key and the analytics skill, then ask it where your money is going
Goal: Run a cost review on your OpenRouter account using your coding agent, the Analytics API, and the openrouter-analytics skill.Outcome: A set of query recipes and agent prompts for digging into your own usage data: which models burn the most, which API keys cause it, and what to repoint or cache.
Then copy this prompt into your agent to run the full cost review.
Analytics queries need a management key from Settings → Management Keys. Regular inference keys get a 403. Management keys can’t make model requests, so the whole workflow is read-only and free, but the data it returns is your org’s full spend breakdown. Treat the key like any other credential.
Before you start
You need:
A management key your agent can read; the skill’s scripts expect it in OPENROUTER_API_KEY (or passed via --api-key)
Node.js with npx if you use the skill’s scripts (they run via npx tsx)
A coding agent (Claude Code, Devin, Cursor, or anything that can run shell commands)
At least a few weeks of real usage on your account, or there’s nothing to analyze
A cost review your agent runs for you. The conversation starts with one question:
how can I reduce costs for next month?
The agent discovers the schema, pulls spend grouped by model, flags the lines where effective price per token is far above your blended rate (total spend divided by total tokens across all traffic, scaled to $ per million tokens), drills into the API keys behind those lines, and hands you actions ranked by dollar impact.We ran this internally and found a preview model burning ~$6.2K/month at roughly 25x the org’s blended rate. One drill-down query later, 98% of it traced to a single batch-pipeline key running a task that never needed a frontier model. The fix was a one-line model swap.The recipes below are the building blocks of that review. Each one is a prompt you can paste into your agent, followed by an under-the-hood look at the query the agent generates and the shape of what comes back.
Setup: install the skill and discover the schema
The skill bundles runnable query scripts so your agent doesn’t hand-write curl calls:
The response lists every metric, dimension, filter operator, and granularity the API currently supports; the meta endpoint reference shows the full shape. Spend metrics (total_usage, usage_*) are in USD. Token metrics are native tokens. cache_hit_rate is a 0 to 1 ratio.Two things to know before reading any output: count metrics can come back as strings (the reference’s example shows them as numbers, so parse defensively and accept both), and metadata.truncated tells you whether the result hit the row limit. If it’s true, your totals are partial; raise limit or narrow the query before drawing conclusions.The API caps queries at 3 dimensions. A third dimension works only when every metric, dimension, and filter in the query is served from the activity rollups (for example model, api_key_id, endpoint, workspace, app, user, variant), or for the token-kind metrics below from their daily rollup (model, api_key_id, workspace, variant, endpoint, provider, data_region, billed_service_tier, long_context); combining three with a dimension or metric that needs raw generations (such as country, or latency percentiles, or provider without a token-kind metric) returns a 400 naming the incompatible fields. For the token-kind metrics, a third dimension also needs a time_range whose start and end are UTC midnights; a range of 31 days or less that starts or ends mid-day returns a 400 naming the off-midnight bound. A fourth dimension always returns a 400. When the agent needs another angle beyond that, it should run separate queries or add a time-axis granularity instead.
Recipe: which models burn the most?
The widest-angle question, and the right one to start with:
Break down my OpenRouter spend by model for last month. Compute theeffective $/Mtok for each model and flag anything far above myblended rate.
Under the hood, your agent generates a query like this (one POST to /api/v1/analytics/query with the management key):
An explicit time_range matters: without one the API defaults to a recent window that may miss the month you asked about.Sample response shape (1 row shown):
From here the agent computes total_usage / tokens_total * 1e6 for each row to get an effective $/Mtok per model (spend over tokens gives dollars per single token; the 1e6 scales it to dollars per million tokens, the unit model pricing is quoted in). Compare each against your blended rate: the same calculation run with no dimensions, so it covers all traffic in the window. A model priced at a large multiple of the blended rate is the strongest signal to chase, but you may not find one; if every model sits near the blended rate, your spend matches your pricing and the levers are elsewhere (cache rate, prompt size, or feature surcharges, covered in the recipes below). In the internal run above, the flagged model had spent $6,185 on 0.25B tokens (6185 / 250000000 * 1e6 ≈ $24.7/Mtok) with a 7.6% cache rate, about 25x that org’s blended rate.
Recipe: which keys drive a model’s cost?
The model row shows where spend concentrates, but the thing you can change is the key, app, or pipeline calling that model. This works whether the model is an outlier or just your biggest fairly-priced line, since the per-key split still shows which workload to optimize:
My spend on google/gemini-3-flash-preview looks too high. Which ofmy API keys is behind it, and what is each one doing?
Under the hood, the agent filters to that model and groups by api_key_id using a filters array on the request body:
api_key_id, app, and workspace resolve to human-readable names in the response, so each row names the key directly. Here’s the row shape, filled with the internal run’s numbers (rounded, key name changed):
The user dimension is the exception: grouping by user returns two fields per row instead of one. user is the account’s display name (or null when the user has no name set), and user_email is their email (null when the user has no email on file). The raw user ID is never returned, so use user_email to cross-reference users against your own records:
In the internal run, this is where the recommendation wrote itself: a batch-pipeline key doing 37K requests at ~$48/Mtok is high-volume, low-complexity work on the wrong model. Repointing it to a cheap production model recovers the whole line item at near-zero risk.A sharper variant of the same prompt skips the model step entirely:
Are any of my keys calling preview or frontier models for high-volumebatch work? Estimate the savings from moving each one to a cheaperproduction model.
Recipe: what did the money actually buy?
total_usage is a single number. The usage_* components split it into what each dollar paid for:
Break my OpenRouter spend into components: raw inference, caching,discounts, web search, and file parsing. Are surcharges or cachewrites a meaningful slice?
Under the hood, the agent queries the component metrics over a time axis:
Each row carries a time-series key named date__<granularity> or created_at__<granularity> (here, created_at__day) depending on which data source the query resolves to, so accept either prefix. The rest of the row is the components in USD. Unused components come back as null, not 0:
usage_upstream: raw inference cost
usage_cache: what caching saved (or cost, for cache writes)
usage_data: discounts, typically negative
usage_web, usage_file: web search and file parsing surcharges
If usage_web or usage_file is a meaningful slice, the fix is gating those features. If usage_cache savings are near zero on a prompt-heavy workload, caching is your lever. And if cache rates are already high, skip the caching advice entirely; in the internal run, caching was near-maxed and the only real lever was model mix.
Separate batch traffic
Batch API requests are recorded with the batch variant, so the variant dimension separates them from synchronous traffic at any time range:
Rows with "variant": "batch" are batch traffic. Filter with { "field": "variant", "operator": "eq", "value": "batch" } to keep only batch requests, or neq to exclude them.
Recipe: what did each token kind cost?
usage_upstream and byok_usage are single numbers too. The token-kind metrics split them by the kind of token they paid for, at the rates actually billed (provider discounts included), not catalog prices:
byok_usage_uncached_input, byok_usage_cache_read, byok_usage_cache_write, byok_usage_output, byok_usage_reasoning: the same split of BYOK inference cost billed by your provider.
uncached_input_tokens and cache_write_tokens: billed token counts for the two input kinds that have no other metric.
token_kind_request_count: the requests that carry the split, to compare against request_count when a row comes back null.
usage_upstream_token_kind_total and byok_usage_token_kind_total: the upstream cost the split adds up to over the same requests, so one query returns the parts and their total.
Reasoning is its own kind only where the provider bills it at its own rate (Gemini and Perplexity today). Everywhere else, including OpenAI, Anthropic, and xAI, reasoning tokens are billed and reported as output, so usage_upstream_reasoning is 0 while reasoning_tokens is not. tokens_completion always includes reasoning tokens, and cached_tokens and reasoning_tokens are provider-reported counts over every request, so don’t divide a per-kind cost by them to get a per-token rate.They are UTC-daily totals over any range up to a year. The same daily totals also serve request_count, tokens_prompt, tokens_completion, reasoning_tokens, cached_tokens, and tokens_total, with the same values as everywhere else, so one query can put the token counts next to the per-kind costs. The daily totals group by model, api_key_id, workspace, variant, endpoint, provider, data_region, billed_service_tier, and long_context, with granularity of day, week, month, or none. Any other dimension or filter (such as app, user, or country), hour or minute granularity, or a timezone reads raw generations instead, which allows at most 31 days; past that the query returns a 400 naming what kept it off the daily totals. A time_range of 31 days or less that starts or ends off a UTC midnight also reads raw generations, however far back it is, so it is not widened to whole days, and with three dimensions such a range returns a 400 instead, so use UTC midnight bounds; a longer one reads the daily totals for every whole UTC day it touches. Without a time_range, the default last 7 days is snapped out to whole UTC days on the daily totals, as it is on the other activity rollups.Group them by pricing qualifiers to see where the rate came from: billed_service_tier (the tier you were billed at, which can differ from the tier that served the request), long_context ("true" when a long-context price tier applied), endpoint (the endpoint UUID, the same endpoint_id that GET /api/v1/activity returns with its provider), provider (the provider that served the request), and data_region (global, or the region the request was routed in, such as europe). A daily pull by model and key with one qualifier looks like this:
They start on 20 November 2026 (UTC), and the first day reads null. The split is recorded as requests are billed, from the release that added these metrics; there is no backfill. The daily totals treat their earliest day as incomplete, because it can hold only part of that day’s traffic, so every metric of a row on or before that day (or of an ungranular query whose range starts on or before it) is null, including request_count. The rule relies on the daily totals being switched on before their first day begins, and on any later backfill filling one unbroken run of days up to and including that first day.
A row is either complete or null. A token-kind metric is null for a row unless every request in it carries the split, that is unless token_kind_request_count equals request_count. Requests whose billed rows can’t be reconciled with what was charged (for example fallback pricing when a provider reports no usage, or batch estimates), or whose charged amount was not recorded, have no split, so a row containing one reads as null rather than as a partial total. Batch requests never carry the split, so keep variant in the grouping (or filter it out) when a model takes both batch and synchronous traffic.
They don’t always add up to usage_upstream. Server tools billed by the provider (such as native web search), Anthropic advisor-model tokens, and other non-token charges belong to no token kind, so the five kinds can sum to less than usage_upstream. Pricing qualifiers are recorded on chat, Responses, and Messages requests; other request types (embeddings, images, batch) group under an empty value.
Recipe: is my workload prompt-heavy?
The token shape decides whether caching or prompt trimming is worth the effort:
Is prompt caching saving me anything? Show prompt vs completiontokens, reasoning tokens, and cache_hit_rate by model, and tell mewhere caching would pay off.
A 20:1 prompt-to-completion ratio points at oversized context, and a large reasoning_tokens share means you’re paying for thinking you may not need. Pair the ratio with cache_hit_rate: prompt-heavy traffic with a low cache rate is the textbook caching win, while prompt-heavy traffic with a high cache rate is already optimized and the lever moves back to model mix.
Recipe: did spend actually change?
Grouping spend over a time axis covers 2 jobs: finding what changed when a bill jumps, and verifying that a fix landed. For the first:
My OpenRouter bill doubled this month. Find what changed, comparingthis month's spend by model and key against last month's.
Under the hood, the agent groups per-key spend over a weekly time axis:
For the bill-doubled case, the key (or model, if the agent groups by model instead) whose weekly line jumped is your culprit. For verification, ask the agent to re-run the same query after a fix ships: a successful repoint shows the key’s weekly total_usage falling off a cliff at the deploy date, like the sample above.Filter values must match what the dimension stores internally, so agents should filter on model slugs (which they know exactly) and group by api_key_id rather than filtering on resolved key names. Save the prompts that worked and re-run them monthly; the copy-prompt at the top chains all of these recipes into one full review.
Use prompt caching where this review showed low cache rates on prompt-heavy traffic.
Was this page helpful?
{"docsConfig":{"theme":"mint","$schema":"https://mintlify.com/docs.json","name":"OpenRouter | Documentation","colors":{"primary":"#7624F4","light":"#7624F4","dark":"#C8FF00"},"logo":{"light":"https://mintcdn.com/openrouter-d02e98a0/ksNSeB_K7gD-BUDh/assets/logo-v2-light.svg?fit=max\u0026auto=format\u0026n=ksNSeB_K7gD-BUDh\u0026q=85\u0026s=690e45873843519efe02bcbe45795dac","dark":"https://mintcdn.com/openrouter-d02e98a0/ksNSeB_K7gD-BUDh/assets/logo-v2-dark.svg?fit=max\u0026auto=format\u0026n=ksNSeB_K7gD-BUDh\u0026q=85\u0026s=8ba2f49b11cd9839c37dd03c62537ebd","href":"https://openrouter.ai"},"favicon":"https://mintcdn.com/openrouter-d02e98a0/n3IgkjAMQY8SgOLX/assets/favicon-v2.svg?fit=max\u0026auto=format\u0026n=n3IgkjAMQY8SgOLX\u0026q=85\u0026s=f49800815fcadee8a07dcc093051c2de","appearance":{"default":"dark"},"background":{"color":{"light":"#FCFCFE","dark":"#03080A"}},"navbar":{"links":[{"href":"https://openrouter.ai/models","label":"Models"},{"href":"https://openrouter.ai/benchmarks","label":"Benchmarks"},{"href":"https://openrouter.ai/chat","label":"Chat"},{"href":"https://openrouter.ai/rankings","label":"Rankings"},{"href":"https://openrouter.ai/apps","label":"Apps"},{"href":"https://openrouter.ai/ori","label":"Ori"},{"href":"https://openrouter.ai/pricing","label":"Pricing"}],"primary":{"type":"button","label":"Sign In","href":"https://openrouter.ai/sign-in"}},"seo":{"metatags":{"canonical":"https://openrouter.ai/docs"}},"fonts":{"family":"Plus Jakarta Sans","weight":400},"integrations":{"ga4":{"measurementId":"G-R8YZRJS2XN"},"posthog":{"apiKey":"phc_7ToS2jDeWBlMu4n2JoNzoA1FnArdKwFMFoHVnAqQ6O1","apiHost":"https://us.i.posthog.com"}},"contextual":{"options":["copy","view","chatgpt","claude","perplexity"],"display":"header"},"styling":{"latex":true},"navigation":{"pages":[]}},"deployment":{"deploymentMetadataValue":{"subdomain":"openrouter-d02e98a0","hasAudioNarration":false,"actualSubdomain":"openrouter-d02e98a0","customDomains":["openrouter.ai"],"gitSource":{"_id":"6ad7fa92692c7c940ff17a31","type":"github","owner":"openrouterteam","repo":"docs","deployBranch":"main","contentDirectory":"","isPrivate":false,"mountPath":"","label":""},"feedback":{"contextualFeedback":{"allowEmail":true},"agentFeedback":true,"issues":false,"thumbs":true},"entitlements":{"AI_CHAT":{"status":"ENABLED"},"REMOVE_BRANDING":{"status":"ENABLED"},"CONTEXTUAL_FEEDBACK":{"status":"ENABLED"},"CODE_SNIPPET_FEEDBACK":{"status":"ENABLED"},"WEB_MCP":{"status":"ENABLED"},"NARRATION":{"status":"ENABLED"}},"buildId":"6afe1bda12636d63bad0484d","clientVersion":"0.0.4205","searchFilterCounts":[{"language":"default","tag":"API Reference","count":355},{"language":"default","tag":"Client SDKs","count":248},{"language":"default","tag":"Docs","count":159},{"language":"default","tag":"Cookbook","count":64},{"language":"default","tag":"Agent SDK","count":42},{"language":"default","tag":"TypeScript SDK","count":41},{"language":"default","tag":"Python SDK","count":40},{"language":"default","tag":"Go SDK","count":39},{"language":"default","tag":"Observability","count":30},{"language":"default","tag":"Features","count":25}]},"assistantConfig":{"enableAskAiButton":false,"enableStarterQuestions":true,"enableContextualStarterQuestions":true,"enableAskAiSearchBar":true},"basePath":"/docs","hasChatPermissions":true,"hasWebMcpPermissions":true,"toggles":[{"name":"kill-isr","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"custom-rbac-roles","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"agent-mintlify-slack","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"agent-integrations-settings","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"rbac-expansion","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automations-pr-grouping","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"custom-prefix","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"analytics-v7","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"user-flows-analytics","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"no-publish-flow","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-onboarding","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"snapshot-serving-live","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"stripTemperatureMiddleware","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"support-platform-pylon-zendesk","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"snapshot-serving","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"instant-rollback","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"granular-notifications-server","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"granular-notifications","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automations-fable","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"fixed-ai-action-pricing","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automation-resolve-merge-conflict","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"scoped-publishing","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"mintql","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"page-level-analytics","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"editor-version-history-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"billing-page-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"full-page-search","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workspace-projects","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"assistant-v5","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"fork-pr-previews","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"revalidation-jwt","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-sso-scim","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"hobby-abuse-scan","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"dashboard-inbox","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-signup","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"turbopuffer-dual-write","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-enabled","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"approve-review-flow","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"editor-metadata","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"sso-setup-deadline-banner","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"turbopuffer-serve","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"live-agents","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"starter-pricing","enabled":true,"variant":{"name":"treatment","enabled":true,"payload":{"type":"string","value":"treatment"},"feature_enabled":true},"impressionData":false},{"name":"llms-txt-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"wildcard-email-auth","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false}],"starterQuestions":[{"_id":"6af272c0af52927e0c2cd6a8","deploymentId":"6a349898c5627d470e2c2ff7","questionText":"How do I use the best free models in OpenRouter?","status":"active","generatedBy":"manual","createdAt":"2026-11-10T01:26:24.136Z","updatedAt":"2026-11-10T01:26:24.136Z","__v":0},{"_id":"6ad8ee98f69c2f90c2a8cc12","deploymentId":"6a349898c5627d470e2c2ff7","questionText":"How do I get started using decision models like Jev?","status":"active","generatedBy":"manual","createdAt":"2026-10-21T16:55:52.120Z","updatedAt":"2026-10-21T16:55:52.120Z","__v":0}],"googleFontsInlined":true},"customCss":[{"filePath":"style.css","content":"/* cyrillic-ext */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 400;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrodmgPn.woff2) format('woff2');\n unicode-range: U+0460-052F, U+1C80-1C8A, U+20B4, U+2DE0-2DFF, U+A640-A69F, U+FE2E-FE2F;\n}\n/* cyrillic */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 400;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrMdmgPn.woff2) format('woff2');\n unicode-range: U+0301, U+0400-045F, U+0490-0491, U+04B0-04B1, U+2116;\n}\n/* symbols2 */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 400;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFg08vz7ehw.woff2) format('woff2');\n unicode-range: U+2000-2001, U+2004-2008, U+200A, U+23B8-23BD, U+2500-259F;\n}\n/* vietnamese */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 400;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrgdmgPn.woff2) format('woff2');\n unicode-range: U+0102-0103, U+0110-0111, U+0128-0129, U+0168-0169, U+01A0-01A1, U+01AF-01B0, U+0300-0301, U+0303-0304, U+0308-0309, U+0323, U+0329, U+1EA0-1EF9, U+20AB;\n}\n/* latin-ext */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 400;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrkdmgPn.woff2) format('woff2');\n unicode-range: U+0100-02BA, U+02BD-02C5, U+02C7-02CC, U+02CE-02D7, U+02DD-02FF, U+0304, U+0308, U+0329, U+1D00-1DBF, U+1E00-1E9F, U+1EF2-1EFF, U+2020, U+20A0-20AB, U+20AD-20C4, U+2113, U+2C60-2C7F, U+A720-A7FF;\n}\n/* latin */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 400;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrcdmg.woff2) format('woff2');\n unicode-range: U+0000-00FF, U+0131, U+0152-0153, U+02BB-02BC, U+02C6, U+02DA, U+02DC, U+0304, U+0308, U+0329, U+2000-206F, U+20AC, U+2122, U+2191, U+2193, U+2212, U+2215, U+FEFF, U+FFFD;\n}\n/* cyrillic-ext */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 500;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrodmgPn.woff2) format('woff2');\n unicode-range: U+0460-052F, U+1C80-1C8A, U+20B4, U+2DE0-2DFF, U+A640-A69F, U+FE2E-FE2F;\n}\n/* cyrillic */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 500;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrMdmgPn.woff2) format('woff2');\n unicode-range: U+0301, U+0400-045F, U+0490-0491, U+04B0-04B1, U+2116;\n}\n/* symbols2 */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 500;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFg08vz7ehw.woff2) format('woff2');\n unicode-range: U+2000-2001, U+2004-2008, U+200A, U+23B8-23BD, U+2500-259F;\n}\n/* vietnamese */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 500;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrgdmgPn.woff2) format('woff2');\n unicode-range: U+0102-0103, U+0110-0111, U+0128-0129, U+0168-0169, U+01A0-01A1, U+01AF-01B0, U+0300-0301, U+0303-0304, U+0308-0309, U+0323, U+0329, U+1EA0-1EF9, U+20AB;\n}\n/* latin-ext */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 500;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrkdmgPn.woff2) format('woff2');\n unicode-range: U+0100-02BA, U+02BD-02C5, U+02C7-02CC, U+02CE-02D7, U+02DD-02FF, U+0304, U+0308, U+0329, U+1D00-1DBF, U+1E00-1E9F, U+1EF2-1EFF, U+2020, U+20A0-20AB, U+20AD-20C4, U+2113, U+2C60-2C7F, U+A720-A7FF;\n}\n/* latin */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 500;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrcdmg.woff2) format('woff2');\n unicode-range: U+0000-00FF, U+0131, U+0152-0153, U+02BB-02BC, U+02C6, U+02DA, U+02DC, U+0304, U+0308, U+0329, U+2000-206F, U+20AC, U+2122, U+2191, U+2193, U+2212, U+2215, U+FEFF, U+FFFD;\n}\n/* cyrillic-ext */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 600;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrodmgPn.woff2) format('woff2');\n unicode-range: U+0460-052F, U+1C80-1C8A, U+20B4, U+2DE0-2DFF, U+A640-A69F, U+FE2E-FE2F;\n}\n/* cyrillic */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 600;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrMdmgPn.woff2) format('woff2');\n unicode-range: U+0301, U+0400-045F, U+0490-0491, U+04B0-04B1, U+2116;\n}\n/* symbols2 */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 600;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFg08vz7ehw.woff2) format('woff2');\n unicode-range: U+2000-2001, U+2004-2008, U+200A, U+23B8-23BD, U+2500-259F;\n}\n/* vietnamese */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 600;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrgdmgPn.woff2) format('woff2');\n unicode-range: U+0102-0103, U+0110-0111, U+0128-0129, U+0168-0169, U+01A0-01A1, U+01AF-01B0, U+0300-0301, U+0303-0304, U+0308-0309, U+0323, U+0329, U+1EA0-1EF9, U+20AB;\n}\n/* latin-ext */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 600;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrkdmgPn.woff2) format('woff2');\n unicode-range: U+0100-02BA, U+02BD-02C5, U+02C7-02CC, U+02CE-02D7, U+02DD-02FF, U+0304, U+0308, U+0329, U+1D00-1DBF, U+1E00-1E9F, U+1EF2-1EFF, U+2020, U+20A0-20AB, U+20AD-20C4, U+2113, U+2C60-2C7F, U+A720-A7FF;\n}\n/* latin */\n@font-face {\n font-family: 'Geist Mono';\n font-style: normal;\n font-weight: 600;\n font-display: swap;\n src: url(/docs/_mintlify/assets/fonts/google/s/geistmono/v6/or3nQ6H-1_WfwkMZI_qYFrcdmg.woff2) format('woff2');\n unicode-range: U+0000-00FF, U+0131, U+0152-0153, U+02BB-02BC, U+02C6, U+02DA, U+02DC, U+0304, U+0308, U+0329, U+2000-206F, U+20AC, U+2122, U+2191, U+2193, U+2212, U+2215, U+FEFF, U+FFFD;\n}\n:root{--inverse: 3 8 10;--inverse-foreground: 252 252 254;--tooltip-background: var(--inverse);--tooltip-foreground: var(--inverse-foreground)}code,pre,pre code,.code-block code,[class*=\"code\"] code{font-family:'Geist Mono', ui-monospace, SFMono-Regular, 'SF Mono', Menlo, Consolas, monospace}::selection{background:rgb(118 36 244 / 0.25)}.dark ::selection,html.dark ::selection{background:rgb(200 255 0 / 0.25);color:inherit}.dark,[data-theme=\"dark\"],html.dark{--primary: 200 255 0;--primary-light: 200 255 0;--primary-dark: 200 255 0;--light-primary: 200 255 0;--primary-foreground: 3 8 10;--inverse: 252 252 254;--inverse-foreground: 3 8 10;--default-border-color: rgb(252 252 254 / 0.14)}#navbar{border-radius:0}#sidebar a[aria-current=\"page\"],nav[aria-label=\"Pages\"] a[aria-current=\"page\"]{border-radius:8px}[data-component-part=\"code-block-root\"]{border-radius:8px}.prose a.link{border-bottom-width:0;text-decoration-line:underline}.prose a.link,a:not([class]){text-decoration-thickness:1px;text-underline-offset:2px;text-decoration-color:color-mix(in srgb, currentColor 40%, transparent)}.prose a.link:hover,a:not([class]):hover{text-decoration-color:currentColor}.nav-tag-pill-text{background:transparent!important;color:rgb(0 0 0 / 0.5)!important;border:1px solid rgb(0 0 0 / 0.2)!important}.dark .nav-tag-pill-text,html.dark .nav-tag-pill-text{background:transparent!important;color:rgb(252 252 254 / 0.5)!important;border:1px solid rgb(252 252 254 / 0.2)!important}.nav-logo{height:1.5rem}.dark nav[aria-label=\"Pages\"] button[aria-expanded] svg.transition-transform,html.dark nav[aria-label=\"Pages\"] button[aria-expanded] svg.transition-transform{color:rgb(252 252 254 / 0.627)}.dark nav[aria-label=\"Pages\"] button[aria-expanded]:hover svg.transition-transform,html.dark nav[aria-label=\"Pages\"] button[aria-expanded]:hover svg.transition-transform{color:rgb(252 252 254)}nav[aria-label=\"Pages\"] button[aria-expanded] svg.transition-transform{width:0.625rem;height:0.625rem}[role=\"tooltip\"][data-side],[data-radix-popper-content-wrapper]>div[data-state=\"delayed-open\"],[data-radix-popper-content-wrapper]>div[data-state=\"instant-open\"]{background-color:rgb(var(--tooltip-background))!important;color:rgb(var(--tooltip-foreground))!important}.callout [data-component-part=\"callout-content\"]>.code-block:first-child{margin-top:0}.callout [data-component-part=\"callout-content\"]>.code-block:last-child{margin-bottom:0}[data-floating-buttons] button{background-color:rgb(241 239 245)}.dark [data-floating-buttons] button,html.dark [data-floating-buttons] button{background-color:rgb(40 38 44)}[data-component-part=\"update-tag\"]{border:1px solid rgb(0 0 0 / 0.2);border-radius:9999px;padding:0.125rem 0.625rem;font-size:0.75rem;line-height:1.25rem;color:rgb(0 0 0 / 0.6)}.dark [data-component-part=\"update-tag\"],html.dark [data-component-part=\"update-tag\"]{border-color:rgb(252 252 254 / 0.2);color:rgb(252 252 254 / 0.6)}#topbar-cta-button svg{display:none}#topbar-cta-button a{padding-right:0.75rem}#topbar-cta-button .bg-primary-dark{background-color:rgb(var(--primary))}.dark #topbar-cta-button .text-white,html.dark #topbar-cta-button .text-white{color:rgb(3 8 10)}.dark #changelog-filters-content button.bg-primary,html.dark #changelog-filters-content button.bg-primary{color:rgb(3 8 10 / 0.9)!important}.changelog-status-badge{margin-left:0.5rem}.changelog-status-badge[data-color=\"gray\"]{border:1px solid rgb(3 8 10 / 0.3)!important}.dark .changelog-status-badge[data-color=\"gray\"],html.dark .changelog-status-badge[data-color=\"gray\"]{border-color:rgb(252 252 254 / 0.4)!important}.mermaid>svg{margin-inline:auto}.support-matrix table{font-size:0.8125rem}.support-matrix td,.support-matrix th{min-width:0;padding-right:0.5rem;white-space:nowrap}.support-matrix th code{padding-left:0.25rem;padding-right:0.25rem;font-size:0.6875rem}.pr-fig,.pr-mini{--pr-ink: #03080a;--pr-muted: rgb(3 8 10 / 0.69);--pr-rule: rgb(3 8 10 / 0.12);--pr-surface: #ffffff;--pr-positive-text: #007544;--pr-negative-text: #bf0024}.dark .pr-fig,.dark .pr-mini{--pr-ink: #fcfcfe;--pr-muted: rgb(252 252 254 / 0.627);--pr-rule: rgb(252 252 254 / 0.14);--pr-surface: #080d0f;--pr-positive-text: #34dfaa;--pr-negative-text: #ff2d55}.pr-fig{margin:1.5rem 0;padding:1rem;overflow-x:auto;container-type:inline-size;border:1px solid var(--pr-rule);border-radius:8px;background:var(--pr-surface)}.pr-fig svg{display:block;width:100%;height:auto}.pr-mini{margin:1rem 0 1.25rem}.pr-mini svg{display:block;max-width:100%;height:auto;margin-inline:auto}.pr-fig svg text,.pr-mini svg text{font-family:inherit;font-size:12px;fill:var(--pr-muted)}.pr-sw{display:inline-block;flex:none;width:12px;height:12px;border-radius:2px}.pr-bg-price{background:#0088fe}.pr-bg-ttft{background:#ffbb28}.pr-bg-speed{background:#7b68ee}.pr-bg-rel{background:#ff2d55}.pr-bg-429{background:#ff8042}.pr-fill-ttft{fill:#ffbb28}.pr-fill-speed{fill:#7b68ee}.pr-fill-rel{fill:#ff2d55}.pr-fill-429{fill:#ff8042}.pr-fig .lbl{font-size:13px;fill:var(--pr-ink)}.pr-fig .num{fill:var(--pr-muted);font-variant-numeric:tabular-nums}.pr-fig .strong{fill:var(--pr-ink);font-weight:600}.pr-fig .halo{paint-order:stroke;stroke:var(--pr-surface);stroke-width:4px;stroke-linejoin:round}.pr-fig .axis{stroke:var(--pr-rule);stroke-width:1}.pr-fig .zone{fill:rgb(0 191 111 / 0.12)}.pr-fig .zone-lbl{font-weight:600;fill:var(--pr-positive-text)}.pr-fig .zone-lbl.muted{fill:var(--pr-muted)}.pr-fig .edge{stroke:#00bf6f;stroke-width:1.5;stroke-dasharray:4 3}.pr-fig .edge-curve{fill:none;stroke:#00bf6f;stroke-width:2;stroke-dasharray:5 3}.pr-fig .curve{fill:none;stroke:var(--pr-ink);stroke-width:2.5}.pr-fig .curve-fill{fill:var(--pr-ink);opacity:0.04}.pr-fig .dot{fill:var(--pr-muted);fill-opacity:0.8;stroke:var(--pr-surface);stroke-width:1.5}.pr-fig .dot-in{fill:#00bf6f;stroke:var(--pr-surface);stroke-width:1.5}.pr-fig .got{fill:none;stroke:var(--pr-ink);stroke-width:2}.pr-fig .got-fill{fill:var(--pr-ink);opacity:0.06}.pr-legend{display:flex;flex-wrap:wrap;justify-content:center;gap:0.375rem 1rem;margin-top:0.75rem;font-size:12px;color:var(--pr-muted)}.pr-legend span{display:inline-flex;align-items:center;gap:0.375rem}.pr-table{margin-top:0.875rem;border-top:1px solid var(--pr-rule);font-size:12px}.pr-row{display:grid;grid-template-columns:76px minmax(0, 1fr) 84px;gap:0.5rem;align-items:center;padding:0.375rem 0;border-bottom:1px solid var(--pr-rule)}.pr-head{color:var(--pr-muted);white-space:nowrap}.pr-sub::before{content:\", \"}.pr-name{font-weight:600;color:var(--pr-ink)}.pr-bar{display:flex;height:10px}.pr-bar i{display:block;height:100%}.pr-share{display:flex;align-items:center;gap:0.375rem}.pr-share i{display:block;height:10px;max-width:54px;border-radius:2px;background:var(--pr-muted);opacity:0.5}.pr-share b{font-weight:600;font-variant-numeric:tabular-nums;color:var(--pr-muted)}@media (max-width: 560px){.pr-row{grid-template-columns:76px minmax(0, 1fr) 32px}.pr-share i{display:none}.pr-head{white-space:normal}.pr-sub{display:block}.pr-sub::before{content:none}}@container (max-width: 480px){.pr-fig svg text{font-size:15px}.pr-fig svg .lbl{font-size:16px}}@container (max-width: 400px){.pr-fig svg text{font-size:17px}.pr-fig svg .lbl{font-size:18px}}.pr-mini-row{display:flex;justify-content:center;align-items:center;gap:0.75rem 1.5rem}.pr-mini-row svg{flex:0 1 auto;min-width:0;margin-inline:0}@media (max-width: 640px){.pr-mini-row{flex-wrap:wrap}}.pr-mini .strong-lbl{fill:var(--pr-ink);font-weight:600}.pr-mini .amt{text-anchor:end;font-variant-numeric:tabular-nums}.pr-mini .op{font-size:20px}.pr-mini .paper{fill:var(--pr-surface);stroke:var(--pr-rule);stroke-width:1}.pr-mini .usual{stroke:var(--pr-muted);stroke-width:1;stroke-dasharray:3 3}.pr-mini .sum-line{stroke:var(--pr-ink);stroke-width:1}.pr-mini .price-pill{fill:#0088fe;opacity:0.12}.pr-mini .pill-text{fill:var(--pr-ink);font-weight:600;font-variant-numeric:tabular-nums}.pr-mini .neg-pill{fill:#ff2d55;opacity:0.12}.pr-mini .neg-text{font-size:13px;fill:var(--pr-negative-text);font-weight:600;font-variant-numeric:tabular-nums}.pr-mini .span-stream{fill:var(--pr-muted);opacity:0.3}.pr-mini .span-total{fill:var(--pr-rule)}.pr-mini .extra{opacity:0.45}.pr-mini .bad{fill:none;stroke:#ff2d55;stroke-width:1.5}.pr-mini .tree{fill:none;stroke:var(--pr-muted);stroke-width:1}.pr-mini .brace{fill:none;stroke:var(--pr-muted);stroke-width:1.25}.pr-mini .big-arrow{fill:none;stroke:var(--pr-ink);stroke-width:3}.pr-mini .big-arrow-head,.pr-mini .user-dot{fill:var(--pr-ink)}.pr-factors ul{list-style:none;padding-left:0}.pr-factors li{padding-left:0!important}.pr-factors li::before{display:none}.pr-factors .pr-sw,h3 .pr-sw{margin-right:0.625rem;transform:translateY(1px)}h3 .pr-sw{width:13px;height:13px;transform:translateY(-2px)}.pr-cols{display:grid;grid-template-columns:repeat(auto-fit, minmax(240px, 1fr));gap:0 1.75rem}"}],"page":{"pagePath":"cookbook/administration/analytics-cost-control","slug":"/cookbook/administration/analytics-cost-control","pageMetadata":{"publish":true,"title":"Control Costs with the Analytics API","description":"Hand your coding agent a management key and the analytics skill, then ask it where your money is going","seoTitle":"Control Costs with the Analytics API | OpenRouter Cookbook","og:site_name":"OpenRouter Documentation","og:title":"Control Costs with the OpenRouter Analytics API","og:description":"Use your coding agent, the Analytics API, and the openrouter-analytics skill to find wasted spend, drill into the keys causing it, and verify the fix.","og:image":"https://openrouter.ai/dynamic-og?title=Control%20Costs%20with%20Analytics\u0026description=Find%20wasted%20spend%20with%20your%20coding%20agent","og:image:width":1200,"og:image:height":630,"twitter:card":"summary_large_image","twitter:site":"@OpenRouter","noindex":false,"nofollow":false,"icon":"chart-line","href":"/cookbook/administration/analytics-cost-control"},"description":{"compiledSource":"\"use strict\";\nconst {jsx: _jsx} = arguments[0];\nconst {useMDXComponents: _provideComponents} = arguments[0];\nfunction _createMdxContent(props) {\n const _components = {\n p: \"p\",\n ..._provideComponents(),\n ...props.components\n };\n return _jsx(_components.p, {\n children: \"Hand your coding agent a management key and the analytics skill, then ask it where your money is going\"\n });\n}\nfunction MDXContent(props = {}) {\n const {wrapper: MDXLayout} = {\n ..._provideComponents(),\n ...props.components\n };\n return MDXLayout ? _jsx(MDXLayout, {\n ...props,\n children: _jsx(_createMdxContent, {\n ...props\n })\n }) : _createMdxContent(props);\n}\nreturn {\n default: MDXContent\n};\n","frontmatter":{},"scope":{}},"pageType":"default","mdx":{"mdxSource":{"compiledSource":"\"use strict\";const{Fragment:_Fragment,jsx:_jsx,jsxs:_jsxs}=arguments[0];const{useMDXComponents:_provideComponents}=arguments[0];const CopyPromptButton=({prompt,buttonLabel=`Copy prompt`})=>{const[isCopied,setIsCopied]=useState(false);const handleClick=()=>{navigator.clipboard.writeText(prompt).then(()=>{setIsCopied(true);setTimeout(()=>setIsCopied(false),2e3)})};return _jsx(`div`,{className:`mint-mt-3`,children:_jsx(`button`,{type:`button`,onClick:handleClick,className:`border-border mint-inline-flex mint-items-center mint-gap-2 mint-rounded-lg mint-border mint-px-3 mint-py-2 mint-font-semibold mint-cursor-pointer mint-border-green-200 dark:mint-border-green-900`,children:isCopied?`Copied`:buttonLabel})})};function _createMdxContent(props){const _components={a:`a`,code:`code`,li:`li`,p:`p`,pre:`pre`,span:`span`,strong:`strong`,ul:`ul`,..._provideComponents(),...props.components},{CodeBlock,Heading,Tip,Warning,_MdxComponentBoundary}=_components;if(!CodeBlock)_missingMdxReference(`CodeBlock`,true);if(!Heading)_missingMdxReference(`Heading`,true);if(!Tip)_missingMdxReference(`Tip`,true);if(!Warning)_missingMdxReference(`Warning`,true);if(!_MdxComponentBoundary)_missingMdxReference(`_MdxComponentBoundary`,true);return _jsxs(_Fragment,{children:[_jsxs(_components.p,{children:[_jsx(_components.strong,{children:`Goal:`}),` Run a cost review on your OpenRouter account using your coding agent, the `,_jsx(_components.a,{href:`/api/api-reference/analytics/query-analytics-data`,children:`Analytics API`}),`, and the `,_jsx(_components.a,{href:`https://github.com/OpenRouterTeam/skills/tree/main/skills/openrouter-analytics`,children:`openrouter-analytics skill`}),`.`]}),`\n`,_jsxs(_components.p,{children:[_jsx(_components.strong,{children:`Outcome:`}),` A set of query recipes and agent prompts for digging into your own usage data: which models burn the most, which API keys cause it, and what to repoint or cache.`]}),`\n`,_jsxs(Tip,{children:[_jsx(_components.p,{children:`Then copy this prompt into your agent to run the full cost review.`}),_jsx(_MdxComponentBoundary,{name:`CopyPromptButton`,children:_jsx(CopyPromptButton,{prompt:`Run a cost review on my OpenRouter account and give me ranked recommendations for reducing spend next month.\n\nUse the openrouter-analytics skill if it's installed (clone https://github.com/OpenRouterTeam/skills and run npm install in skills/openrouter-analytics/scripts). Otherwise call the API directly. Either way, read these source-of-truth docs for the current query schema, metrics, and dimensions before writing any query:\n- Query endpoint reference: https://openrouter.ai/docs/api/api-reference/analytics/query-analytics-data\n- Skill with runnable query scripts: https://github.com/OpenRouterTeam/skills/tree/main/skills/openrouter-analytics\n\nAuth: the Analytics API needs an OpenRouter management key (regular inference keys get a 403). Look for one in my environment or secret store first; the skill's scripts read it from OPENROUTER_API_KEY or a --api-key flag. If you can't find one, stop and ask me to create one at https://openrouter.ai/settings/management-keys instead of guessing.\n\nMethod:\n1. Fetch /api/v1/analytics/meta and use only metrics, dimensions, and operators it actually returns.\n2. Inspect this codebase (and any sibling services you can see) to map API key names and apps in the analytics data to the pipelines, jobs, or features that call them, so recommendations name the actual code to change.\n3. Establish the baseline: total spend, tokens, and requests over the last full month, plus the blended $/Mtok rate (total_usage / tokens_total * 1e6).\n4. Break spend down by model with cache_hit_rate and tokens. Flag any model whose effective $/Mtok is a large multiple of the blended rate, especially preview or frontier models on high-volume traffic.\n5. For each flagged model, drill into api_key_id (and app or user if populated) with a filter on that model to find who's responsible.\n6. Decompose spend with the usage_* component metrics to see if money is going to upstream inference, web search, file parsing, or being saved by caching and discounts.\n7. Check prompt-heaviness (tokens_prompt vs tokens_completion) and reasoning_tokens share to decide whether caching or prompt trimming is even worth recommending.\n\nFor every recommendation, state the monthly dollar impact, the specific key or pipeline to change, the suggested replacement (cheaper model, caching, prompt change), and the risk. Order by impact. Skip generic advice that my data doesn't support. Mind that latency/throughput metrics and the generation_id dimension only work for time ranges up to 31 days, and check metadata.truncated on every response before summing rows.\n\nFinish with the exact queries to re-run next month to verify each fix worked.`})})]}),`\n`,_jsx(Warning,{children:_jsxs(_components.p,{children:[`Analytics queries need a `,_jsx(_components.strong,{children:`management key`}),` from `,_jsx(_components.a,{href:`https://openrouter.ai/settings/management-keys`,children:`Settings → Management Keys`}),`. Regular inference keys get a 403. Management keys can’t make model requests, so the whole workflow is read-only and free, but the data it returns is your org’s full spend breakdown. Treat the key like any other credential.`]})}),`\n`,_jsx(Heading,{level:`2`,id:`before-you-start`,children:`Before you start`}),`\n`,_jsx(_components.p,{children:`You need:`}),`\n`,_jsxs(_components.ul,{children:[`\n`,_jsxs(_components.li,{children:[`A management key your agent can read; the skill’s scripts expect it in `,_jsx(_components.code,{children:`OPENROUTER_API_KEY`}),` (or passed via `,_jsx(_components.code,{children:`--api-key`}),`)`]}),`\n`,_jsxs(_components.li,{children:[`Node.js with `,_jsx(_components.code,{children:`npx`}),` if you use the skill’s scripts (they run via `,_jsx(_components.code,{children:`npx tsx`}),`)`]}),`\n`,_jsx(_components.li,{children:`A coding agent (Claude Code, Devin, Cursor, or anything that can run shell commands)`}),`\n`,_jsx(_components.li,{children:`At least a few weeks of real usage on your account, or there’s nothing to analyze`}),`\n`]}),`\n`,_jsx(_components.p,{children:`Use these references for exact schemas:`}),`\n`,_jsxs(_components.ul,{children:[`\n`,_jsx(_components.li,{children:_jsx(_components.a,{href:`/api/api-reference/analytics/query-analytics-data`,children:`Query analytics endpoint`})}),`\n`,_jsx(_components.li,{children:_jsx(_components.a,{href:`/api/api-reference/analytics/get-available-analytics-metrics-and-dimensions`,children:`Get analytics metadata`})}),`\n`,_jsx(_components.li,{children:_jsx(_components.a,{href:`/guides/overview/auth/management-api-keys`,children:`Management API keys`})}),`\n`,_jsx(_components.li,{children:_jsx(_components.a,{href:`https://github.com/OpenRouterTeam/skills/tree/main/skills/openrouter-analytics`,children:`openrouter-analytics skill`})}),`\n`]}),`\n`,_jsx(Heading,{level:`2`,id:`what-you’re-building`,children:`What you’re building`}),`\n`,_jsx(_components.p,{children:`A cost review your agent runs for you. The conversation starts with one question:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`2`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`2`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`how can I reduce costs for next month?`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})}),`\n`,_jsx(_components.p,{children:`The agent discovers the schema, pulls spend grouped by model, flags the lines where effective price per token is far above your blended rate (total spend divided by total tokens across all traffic, scaled to $ per million tokens), drills into the API keys behind those lines, and hands you actions ranked by dollar impact.`}),`\n`,_jsx(_components.p,{children:`We ran this internally and found a preview model burning ~$6.2K/month at roughly 25x the org’s blended rate. One drill-down query later, 98% of it traced to a single batch-pipeline key running a task that never needed a frontier model. The fix was a one-line model swap.`}),`\n`,_jsx(_components.p,{children:`The recipes below are the building blocks of that review. Each one is a prompt you can paste into your agent, followed by an under-the-hood look at the query the agent generates and the shape of what comes back.`}),`\n`,_jsx(Heading,{level:`2`,id:`setup-install-the-skill-and-discover-the-schema`,children:`Setup: install the skill and discover the schema`}),`\n`,_jsxs(_components.p,{children:[`The skill bundles runnable query scripts so your agent doesn’t hand-write `,_jsx(_components.code,{children:`curl`}),` calls:`]}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`2`,language:`shellscript`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`shellscript`,children:_jsxs(_components.code,{language:`shellscript`,numberOfLines:`2`,children:[_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#953800`,\"--shiki-dark\":`#DCDCAA`},children:`git`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:` clone`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:` https://github.com/OpenRouterTeam/skills`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#DCDCAA`},children:`cd`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:` skills/skills/openrouter-analytics/scripts`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` \u0026\u0026 `}),_jsx(_components.span,{style:{color:`#953800`,\"--shiki-dark\":`#DCDCAA`},children:`npm`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:` install`})]}),`\n`]})})}),`\n`,_jsx(_components.p,{children:`Schema discovery comes first. Metrics and dimensions evolve, so query what’s actually there instead of trusting a doc snapshot:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`1`,language:`shellscript`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`shellscript`,children:_jsxs(_components.code,{language:`shellscript`,numberOfLines:`1`,children:[_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#953800`,\"--shiki-dark\":`#DCDCAA`},children:`npx`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:` tsx`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:` discover-schema.ts`})]}),`\n`]})})}),`\n`,_jsx(_components.p,{children:`Or hit the endpoint directly:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`2`,language:`shellscript`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`shellscript`,children:_jsxs(_components.code,{language:`shellscript`,numberOfLines:`2`,children:[_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#953800`,\"--shiki-dark\":`#DCDCAA`},children:`curl`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:` https://openrouter.ai/api/v1/analytics/meta`}),_jsx(_components.span,{style:{color:`#CF222E`,\"--shiki-dark\":`#D7BA7D`},children:` \\\\`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#569CD6`},children:` -H`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:` \"Authorization: Bearer `}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#9CDCFE`},children:`$OPENROUTER_API_KEY`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"`})]}),`\n`]})})}),`\n`,_jsxs(_components.p,{children:[`The response lists every metric, dimension, filter operator, and granularity the API currently supports; the `,_jsx(_components.a,{href:`https://openrouter.ai/docs/api/api-reference/analytics/get-available-analytics-metrics-and-dimensions`,children:`meta endpoint reference`}),` shows the full shape. Spend metrics (`,_jsx(_components.code,{children:`total_usage`}),`, `,_jsx(_components.code,{children:`usage_*`}),`) are in USD. Token metrics are native tokens. `,_jsx(_components.code,{children:`cache_hit_rate`}),` is a 0 to 1 ratio.`]}),`\n`,_jsxs(_components.p,{children:[`Two things to know before reading any output: count metrics can come back as `,_jsx(_components.strong,{children:`strings`}),` (the reference’s example shows them as numbers, so parse defensively and accept both), and `,_jsx(_components.code,{children:`metadata.truncated`}),` tells you whether the result hit the row limit. If it’s `,_jsx(_components.code,{children:`true`}),`, your totals are partial; raise `,_jsx(_components.code,{children:`limit`}),` or narrow the query before drawing conclusions.`]}),`\n`,_jsxs(_components.p,{children:[`The API caps queries at 3 dimensions. A third dimension works only when every metric, dimension, and filter in the query is served from the activity rollups (for example `,_jsx(_components.code,{children:`model`}),`, `,_jsx(_components.code,{children:`api_key_id`}),`, `,_jsx(_components.code,{children:`endpoint`}),`, `,_jsx(_components.code,{children:`workspace`}),`, `,_jsx(_components.code,{children:`app`}),`, `,_jsx(_components.code,{children:`user`}),`, `,_jsx(_components.code,{children:`variant`}),`), or for the token-kind metrics below from their daily rollup (`,_jsx(_components.code,{children:`model`}),`, `,_jsx(_components.code,{children:`api_key_id`}),`, `,_jsx(_components.code,{children:`workspace`}),`, `,_jsx(_components.code,{children:`variant`}),`, `,_jsx(_components.code,{children:`endpoint`}),`, `,_jsx(_components.code,{children:`provider`}),`, `,_jsx(_components.code,{children:`data_region`}),`, `,_jsx(_components.code,{children:`billed_service_tier`}),`, `,_jsx(_components.code,{children:`long_context`}),`); combining three with a dimension or metric that needs raw generations (such as `,_jsx(_components.code,{children:`country`}),`, or latency percentiles, or `,_jsx(_components.code,{children:`provider`}),` without a token-kind metric) returns a 400 naming the incompatible fields. For the token-kind metrics, a third dimension also needs a `,_jsx(_components.code,{children:`time_range`}),` whose start and end are UTC midnights; a range of 31 days or less that starts or ends mid-day returns a 400 naming the off-midnight bound. A fourth dimension always returns a 400. When the agent needs another angle beyond that, it should run separate queries or add a time-axis `,_jsx(_components.code,{children:`granularity`}),` instead.`]}),`\n`,_jsx(Heading,{level:`2`,id:`recipe-which-models-burn-the-most`,children:`Recipe: which models burn the most?`}),`\n`,_jsx(_components.p,{children:`The widest-angle question, and the right one to start with:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`Break down my OpenRouter spend by model for last month. Compute the`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`effective $/Mtok for each model and flag anything far above my`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`blended rate.`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})}),`\n`,_jsxs(_components.p,{children:[`Under the hood, your agent generates a query like this (one POST to `,_jsx(_components.code,{children:`/api/v1/analytics/query`}),` with the management key):`]}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`7`,language:`json`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`json`,children:_jsxs(_components.code,{language:`json`,numberOfLines:`7`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"metrics\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"total_usage\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"request_count\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"tokens_total\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"cache_hit_rate\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"dimensions\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"model\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"order_by\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"field\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"total_usage\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"direction\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"desc\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` },`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"time_range\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"start\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-05-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"end\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-06-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` },`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"limit\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`10`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`}`})}),`\n`]})})}),`\n`,_jsxs(_components.p,{children:[`An explicit `,_jsx(_components.code,{children:`time_range`}),` matters: without one the API defaults to a recent window that may miss the month you asked about.`]}),`\n`,_jsx(_components.p,{children:`Sample response shape (1 row shown):`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`14`,language:`json`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`json`,children:_jsxs(_components.code,{language:`json`,numberOfLines:`14`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"data\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: {`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"data\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` {`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"model\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"openrouter/owl-alpha\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`,`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"total_usage\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`0.005`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`,`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"request_count\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"6\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`,`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"tokens_total\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"6331\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`,`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"cache_hit_rate\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`0.17561310238381067`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` ],`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"metadata\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"query_time_ms\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`17`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"row_count\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`1`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"truncated\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#569CD6`},children:`false`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`}`})}),`\n`]})})}),`\n`,_jsxs(_components.p,{children:[`From here the agent computes `,_jsx(_components.code,{children:`total_usage / tokens_total * 1e6`}),` for each row to get an effective $/Mtok per model (spend over tokens gives dollars per single token; the 1e6 scales it to dollars per million tokens, the unit model pricing is quoted in). Compare each against your blended rate: the same calculation run with no dimensions, so it covers all traffic in the window. A model priced at a large multiple of the blended rate is the strongest signal to chase, but you may not find one; if every model sits near the blended rate, your spend matches your pricing and the levers are elsewhere (cache rate, prompt size, or feature surcharges, covered in the recipes below). In the internal run above, the flagged model had spent $6,185 on 0.25B tokens (6185 / 250000000 * 1e6 ≈ $24.7/Mtok) with a 7.6% cache rate, about 25x that org’s blended rate.`]}),`\n`,_jsx(Heading,{level:`2`,id:`recipe-which-keys-drive-a-model’s-cost`,children:`Recipe: which keys drive a model’s cost?`}),`\n`,_jsx(_components.p,{children:`The model row shows where spend concentrates, but the thing you can change is the key, app, or pipeline calling that model. This works whether the model is an outlier or just your biggest fairly-priced line, since the per-key split still shows which workload to optimize:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`3`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`3`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`My spend on google/gemini-3-flash-preview looks too high. Which of`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`my API keys is behind it, and what is each one doing?`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})}),`\n`,_jsxs(_components.p,{children:[`Under the hood, the agent filters to that model and groups by `,_jsx(_components.code,{children:`api_key_id`}),` using a `,_jsx(_components.code,{children:`filters`}),` array on the request body:`]}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`10`,language:`json`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`json`,children:_jsxs(_components.code,{language:`json`,numberOfLines:`10`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"metrics\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"total_usage\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"tokens_total\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"request_count\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"dimensions\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"api_key_id\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"filters\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"field\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"model\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"operator\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"eq\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"value\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"google/gemini-3-flash-preview\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` ],`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"order_by\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"field\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"total_usage\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"direction\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"desc\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` },`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"time_range\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"start\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-05-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"end\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-06-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` },`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"limit\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`10`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`}`})}),`\n`]})})}),`\n`,_jsxs(_components.p,{children:[_jsx(_components.code,{children:`api_key_id`}),`, `,_jsx(_components.code,{children:`app`}),`, and `,_jsx(_components.code,{children:`workspace`}),` resolve to human-readable names in the response, so each row names the key directly. Here’s the row shape, filled with the internal run’s numbers (rounded, key name changed):`]}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`6`,language:`json`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`json`,children:_jsxs(_components.code,{language:`json`,numberOfLines:`6`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"api_key_id\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"batch-pipeline\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`,`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"total_usage\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`6067.0`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`,`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"tokens_total\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"127000000\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`,`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"request_count\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"37000\"`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`}`})}),`\n`]})})}),`\n`,_jsxs(_components.p,{children:[`The `,_jsx(_components.code,{children:`user`}),` dimension is the exception: grouping by `,_jsx(_components.code,{children:`user`}),` returns `,_jsx(_components.strong,{children:`two`}),` fields per row instead of one. `,_jsx(_components.code,{children:`user`}),` is the account’s display name (or `,_jsx(_components.code,{children:`null`}),` when the user has no name set), and `,_jsx(_components.code,{children:`user_email`}),` is their email (`,_jsx(_components.code,{children:`null`}),` when the user has no email on file). The raw user ID is never returned, so use `,_jsx(_components.code,{children:`user_email`}),` to cross-reference users against your own records:`]}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`3`,language:`json`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`json`,children:_jsxs(_components.code,{language:`json`,numberOfLines:`3`,children:[_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{ `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"user\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"Ada Lovelace\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"user_email\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"ada@example.com\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"request_count\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`42`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{ `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"user\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#569CD6`},children:`null`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"user_email\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"grace@example.com\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"request_count\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`7`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{ `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"user\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"Alan Turing\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"user_email\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#569CD6`},children:`null`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"request_count\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`3`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})]}),`\n`]})})}),`\n`,_jsx(_components.p,{children:`In the internal run, this is where the recommendation wrote itself: a batch-pipeline key doing 37K requests at ~$48/Mtok is high-volume, low-complexity work on the wrong model. Repointing it to a cheap production model recovers the whole line item at near-zero risk.`}),`\n`,_jsx(_components.p,{children:`A sharper variant of the same prompt skips the model step entirely:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`Are any of my keys calling preview or frontier models for high-volume`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`batch work? Estimate the savings from moving each one to a cheaper`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`production model.`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})}),`\n`,_jsx(Heading,{level:`2`,id:`recipe-what-did-the-money-actually-buy`,children:`Recipe: what did the money actually buy?`}),`\n`,_jsxs(_components.p,{children:[_jsx(_components.code,{children:`total_usage`}),` is a single number. The `,_jsx(_components.code,{children:`usage_*`}),` components split it into what each dollar paid for:`]}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`Break my OpenRouter spend into components: raw inference, caching,`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`discounts, web search, and file parsing. Are surcharges or cache`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`writes a meaningful slice?`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})}),`\n`,_jsx(_components.p,{children:`Under the hood, the agent queries the component metrics over a time axis:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`5`,language:`json`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`json`,children:_jsxs(_components.code,{language:`json`,numberOfLines:`5`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"metrics\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"usage_upstream\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"usage_cache\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"usage_data\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"usage_web\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"usage_file\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"granularity\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"day\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`,`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"time_range\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"start\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-05-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"end\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-06-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`}`})}),`\n`]})})}),`\n`,_jsxs(_components.p,{children:[`Each row carries a time-series key named `,_jsx(_components.code,{children:`date__\u003cgranularity>`}),` or `,_jsx(_components.code,{children:`created_at__\u003cgranularity>`}),` (here, `,_jsx(_components.code,{children:`created_at__day`}),`) depending on which data source the query resolves to, so accept either prefix. The rest of the row is the components in USD. Unused components come back as `,_jsx(_components.code,{children:`null`}),`, not 0:`]}),`\n`,_jsxs(_components.ul,{children:[`\n`,_jsxs(_components.li,{children:[_jsx(_components.code,{children:`usage_upstream`}),`: raw inference cost`]}),`\n`,_jsxs(_components.li,{children:[_jsx(_components.code,{children:`usage_cache`}),`: what caching saved (or cost, for cache writes)`]}),`\n`,_jsxs(_components.li,{children:[_jsx(_components.code,{children:`usage_data`}),`: discounts, typically negative`]}),`\n`,_jsxs(_components.li,{children:[_jsx(_components.code,{children:`usage_web`}),`, `,_jsx(_components.code,{children:`usage_file`}),`: web search and file parsing surcharges`]}),`\n`]}),`\n`,_jsxs(_components.p,{children:[`If `,_jsx(_components.code,{children:`usage_web`}),` or `,_jsx(_components.code,{children:`usage_file`}),` is a meaningful slice, the fix is gating those features. If `,_jsx(_components.code,{children:`usage_cache`}),` savings are near zero on a prompt-heavy workload, caching is your lever. And if cache rates are already high, skip the caching advice entirely; in the internal run, caching was near-maxed and the only real lever was model mix.`]}),`\n`,_jsx(Heading,{level:`3`,id:`separate-batch-traffic`,children:`Separate batch traffic`}),`\n`,_jsxs(_components.p,{children:[`Batch API requests are recorded with the `,_jsx(_components.code,{children:`batch`}),` variant, so the `,_jsx(_components.code,{children:`variant`}),` dimension separates them from synchronous traffic at any time range:`]}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`6`,language:`json`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`json`,children:_jsxs(_components.code,{language:`json`,numberOfLines:`6`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"metrics\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"total_usage\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"usage_upstream\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"request_count\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"dimensions\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"variant\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"granularity\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"day\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`,`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"time_range\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"start\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-05-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"end\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-06-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`}`})}),`\n`]})})}),`\n`,_jsxs(_components.p,{children:[`Rows with `,_jsx(_components.code,{children:`\"variant\": \"batch\"`}),` are batch traffic. Filter with `,_jsx(_components.code,{children:`{ \"field\": \"variant\", \"operator\": \"eq\", \"value\": \"batch\" }`}),` to keep only batch requests, or `,_jsx(_components.code,{children:`neq`}),` to exclude them.`]}),`\n`,_jsx(Heading,{level:`2`,id:`recipe-what-did-each-token-kind-cost`,children:`Recipe: what did each token kind cost?`}),`\n`,_jsxs(_components.p,{children:[_jsx(_components.code,{children:`usage_upstream`}),` and `,_jsx(_components.code,{children:`byok_usage`}),` are single numbers too. The token-kind metrics split them by the kind of token they paid for, at the rates actually billed (provider discounts included), not catalog prices:`]}),`\n`,_jsxs(_components.ul,{children:[`\n`,_jsxs(_components.li,{children:[_jsx(_components.code,{children:`usage_upstream_uncached_input`}),`, `,_jsx(_components.code,{children:`usage_upstream_cache_read`}),`, `,_jsx(_components.code,{children:`usage_upstream_cache_write`}),`, `,_jsx(_components.code,{children:`usage_upstream_output`}),`, `,_jsx(_components.code,{children:`usage_upstream_reasoning`}),`: upstream inference cost OpenRouter paid, in USD.`]}),`\n`,_jsxs(_components.li,{children:[_jsx(_components.code,{children:`byok_usage_uncached_input`}),`, `,_jsx(_components.code,{children:`byok_usage_cache_read`}),`, `,_jsx(_components.code,{children:`byok_usage_cache_write`}),`, `,_jsx(_components.code,{children:`byok_usage_output`}),`, `,_jsx(_components.code,{children:`byok_usage_reasoning`}),`: the same split of BYOK inference cost billed by your provider.`]}),`\n`,_jsxs(_components.li,{children:[_jsx(_components.code,{children:`uncached_input_tokens`}),` and `,_jsx(_components.code,{children:`cache_write_tokens`}),`: billed token counts for the two input kinds that have no other metric.`]}),`\n`,_jsxs(_components.li,{children:[_jsx(_components.code,{children:`token_kind_request_count`}),`: the requests that carry the split, to compare against `,_jsx(_components.code,{children:`request_count`}),` when a row comes back `,_jsx(_components.code,{children:`null`}),`.`]}),`\n`,_jsxs(_components.li,{children:[_jsx(_components.code,{children:`usage_upstream_token_kind_total`}),` and `,_jsx(_components.code,{children:`byok_usage_token_kind_total`}),`: the upstream cost the split adds up to over the same requests, so one query returns the parts and their total.`]}),`\n`]}),`\n`,_jsxs(_components.p,{children:[`Reasoning is its own kind only where the provider bills it at its own rate (Gemini and Perplexity today). Everywhere else, including OpenAI, Anthropic, and xAI, reasoning tokens are billed and reported as output, so `,_jsx(_components.code,{children:`usage_upstream_reasoning`}),` is `,_jsx(_components.code,{children:`0`}),` while `,_jsx(_components.code,{children:`reasoning_tokens`}),` is not. `,_jsx(_components.code,{children:`tokens_completion`}),` always includes reasoning tokens, and `,_jsx(_components.code,{children:`cached_tokens`}),` and `,_jsx(_components.code,{children:`reasoning_tokens`}),` are provider-reported counts over every request, so don’t divide a per-kind cost by them to get a per-token rate.`]}),`\n`,_jsxs(_components.p,{children:[`They are UTC-daily totals over any range up to a year. The same daily totals also serve `,_jsx(_components.code,{children:`request_count`}),`, `,_jsx(_components.code,{children:`tokens_prompt`}),`, `,_jsx(_components.code,{children:`tokens_completion`}),`, `,_jsx(_components.code,{children:`reasoning_tokens`}),`, `,_jsx(_components.code,{children:`cached_tokens`}),`, and `,_jsx(_components.code,{children:`tokens_total`}),`, with the same values as everywhere else, so one query can put the token counts next to the per-kind costs. The daily totals group by `,_jsx(_components.code,{children:`model`}),`, `,_jsx(_components.code,{children:`api_key_id`}),`, `,_jsx(_components.code,{children:`workspace`}),`, `,_jsx(_components.code,{children:`variant`}),`, `,_jsx(_components.code,{children:`endpoint`}),`, `,_jsx(_components.code,{children:`provider`}),`, `,_jsx(_components.code,{children:`data_region`}),`, `,_jsx(_components.code,{children:`billed_service_tier`}),`, and `,_jsx(_components.code,{children:`long_context`}),`, with `,_jsx(_components.code,{children:`granularity`}),` of `,_jsx(_components.code,{children:`day`}),`, `,_jsx(_components.code,{children:`week`}),`, `,_jsx(_components.code,{children:`month`}),`, or none. Any other dimension or filter (such as `,_jsx(_components.code,{children:`app`}),`, `,_jsx(_components.code,{children:`user`}),`, or `,_jsx(_components.code,{children:`country`}),`), `,_jsx(_components.code,{children:`hour`}),` or `,_jsx(_components.code,{children:`minute`}),` granularity, or a timezone reads raw generations instead, which allows at most 31 days; past that the query returns a 400 naming what kept it off the daily totals. A `,_jsx(_components.code,{children:`time_range`}),` of 31 days or less that starts or ends off a UTC midnight also reads raw generations, however far back it is, so it is not widened to whole days, and with three dimensions such a range returns a 400 instead, so use UTC midnight bounds; a longer one reads the daily totals for every whole UTC day it touches. Without a `,_jsx(_components.code,{children:`time_range`}),`, the default last 7 days is snapped out to whole UTC days on the daily totals, as it is on the other activity rollups.`]}),`\n`,_jsxs(_components.p,{children:[`Group them by pricing qualifiers to see where the rate came from: `,_jsx(_components.code,{children:`billed_service_tier`}),` (the tier you were billed at, which can differ from the tier that served the request), `,_jsx(_components.code,{children:`long_context`}),` (`,_jsx(_components.code,{children:`\"true\"`}),` when a long-context price tier applied), `,_jsx(_components.code,{children:`endpoint`}),` (the endpoint UUID, the same `,_jsx(_components.code,{children:`endpoint_id`}),` that `,_jsx(_components.a,{href:`/api/api-reference/analytics/get-user-activity-grouped-by-endpoint`,children:`GET /api/v1/activity`}),` returns with its provider), `,_jsx(_components.code,{children:`provider`}),` (the provider that served the request), and `,_jsx(_components.code,{children:`data_region`}),` (`,_jsx(_components.code,{children:`global`}),`, or the region the request was routed in, such as `,_jsx(_components.code,{children:`europe`}),`). A daily pull by model and key with one qualifier looks like this:`]}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`6`,language:`json`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`json`,children:_jsxs(_components.code,{language:`json`,numberOfLines:`6`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"metrics\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"usage_upstream_uncached_input\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"usage_upstream_cache_read\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"usage_upstream_cache_write\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"usage_upstream_output\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"usage_upstream_reasoning\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"dimensions\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"model\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"api_key_id\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"billed_service_tier\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"granularity\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"day\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`,`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"time_range\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"start\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-12-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"end\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2027-01-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`}`})}),`\n`]})})}),`\n`,_jsx(_components.p,{children:`Three things to know before summing them:`}),`\n`,_jsxs(_components.ul,{children:[`\n`,_jsxs(_components.li,{children:[_jsxs(_components.strong,{children:[`They start on 20 November 2026 (UTC), and the first day reads `,_jsx(_components.code,{children:`null`}),`.`]}),` The split is recorded as requests are billed, from the release that added these metrics; there is no backfill. The daily totals treat their earliest day as incomplete, because it can hold only part of that day’s traffic, so every metric of a row on or before that day (or of an ungranular query whose range starts on or before it) is `,_jsx(_components.code,{children:`null`}),`, including `,_jsx(_components.code,{children:`request_count`}),`. The rule relies on the daily totals being switched on before their first day begins, and on any later backfill filling one unbroken run of days up to and including that first day.`]}),`\n`,_jsxs(_components.li,{children:[_jsxs(_components.strong,{children:[`A row is either complete or `,_jsx(_components.code,{children:`null`}),`.`]}),` A token-kind metric is `,_jsx(_components.code,{children:`null`}),` for a row unless every request in it carries the split, that is unless `,_jsx(_components.code,{children:`token_kind_request_count`}),` equals `,_jsx(_components.code,{children:`request_count`}),`. Requests whose billed rows can’t be reconciled with what was charged (for example fallback pricing when a provider reports no usage, or batch estimates), or whose charged amount was not recorded, have no split, so a row containing one reads as `,_jsx(_components.code,{children:`null`}),` rather than as a partial total. Batch requests never carry the split, so keep `,_jsx(_components.code,{children:`variant`}),` in the grouping (or filter it out) when a model takes both batch and synchronous traffic.`]}),`\n`,_jsxs(_components.li,{children:[_jsxs(_components.strong,{children:[`They don’t always add up to `,_jsx(_components.code,{children:`usage_upstream`}),`.`]}),` Server tools billed by the provider (such as native web search), Anthropic advisor-model tokens, and other non-token charges belong to no token kind, so the five kinds can sum to less than `,_jsx(_components.code,{children:`usage_upstream`}),`. Pricing qualifiers are recorded on chat, Responses, and Messages requests; other request types (embeddings, images, batch) group under an empty value.`]}),`\n`]}),`\n`,_jsx(Heading,{level:`2`,id:`recipe-is-my-workload-prompt-heavy`,children:`Recipe: is my workload prompt-heavy?`}),`\n`,_jsx(_components.p,{children:`The token shape decides whether caching or prompt trimming is worth the effort:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`4`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`4`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`Is prompt caching saving me anything? Show prompt vs completion`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`tokens, reasoning tokens, and cache_hit_rate by model, and tell me`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`where caching would pay off.`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})}),`\n`,_jsx(_components.p,{children:`Under the hood:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`7`,language:`json`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`json`,children:_jsxs(_components.code,{language:`json`,numberOfLines:`7`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"metrics\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"tokens_prompt\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"tokens_completion\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"reasoning_tokens\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"cache_hit_rate\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"dimensions\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"model\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"order_by\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"field\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"tokens_prompt\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"direction\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"desc\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` },`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"time_range\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"start\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-05-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"end\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-06-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` },`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"limit\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`10`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`}`})}),`\n`]})})}),`\n`,_jsxs(_components.p,{children:[`A 20:1 prompt-to-completion ratio points at oversized context, and a large `,_jsx(_components.code,{children:`reasoning_tokens`}),` share means you’re paying for thinking you may not need. Pair the ratio with `,_jsx(_components.code,{children:`cache_hit_rate`}),`: prompt-heavy traffic with a low cache rate is the textbook caching win, while prompt-heavy traffic with a high cache rate is already optimized and the lever moves back to model mix.`]}),`\n`,_jsx(Heading,{level:`2`,id:`recipe-did-spend-actually-change`,children:`Recipe: did spend actually change?`}),`\n`,_jsx(_components.p,{children:`Grouping spend over a time axis covers 2 jobs: finding what changed when a bill jumps, and verifying that a fix landed. For the first:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`3`,language:`text`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`text`,children:_jsxs(_components.code,{language:`text`,numberOfLines:`3`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`My OpenRouter bill doubled this month. Find what changed, comparing`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{children:`this month's spend by model and key against last month's.`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{})})]})})}),`\n`,_jsx(_components.p,{children:`Under the hood, the agent groups per-key spend over a weekly time axis:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`6`,language:`json`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`json`,children:_jsxs(_components.code,{language:`json`,numberOfLines:`6`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"metrics\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"total_usage\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"dimensions\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"api_key_id\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`],`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"granularity\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"week\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`,`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"time_range\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"start\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-05-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"end\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-07-01T00:00:00Z\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`}`})}),`\n`]})})}),`\n`,_jsx(_components.p,{children:`Sample output:`}),`\n`,_jsx(CodeBlock,{filename:``,lines:`true`,numberOfLines:`9`,language:`json`,children:_jsx(_components.pre,{className:`shiki shiki-themes github-light-default dark-plus`,style:{backgroundColor:`#ffffff`,\"--shiki-dark-bg\":`#0B0C0E`,color:`#1f2328`,\"--shiki-dark\":`#D4D4D4`},language:`json`,children:_jsxs(_components.code,{language:`json`,numberOfLines:`9`,children:[_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`{`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"data\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: {`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"data\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: [`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"date__week\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-06-07\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"api_key_id\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"batch-pipeline\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"total_usage\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`11.2`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` },`})]}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"date__week\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"2026-05-31\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"api_key_id\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0A3069`,\"--shiki-dark\":`#CE9178`},children:`\"batch-pipeline\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"total_usage\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`1402.5`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` ],`})}),`\n`,_jsxs(_components.span,{className:`line`,children:[_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:` \"metadata\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: { `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"query_time_ms\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`7`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"row_count\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#B5CEA8`},children:`2`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`, `}),_jsx(_components.span,{style:{color:`#116329`,\"--shiki-dark\":`#9CDCFE`},children:`\"truncated\"`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`: `}),_jsx(_components.span,{style:{color:`#0550AE`,\"--shiki-dark\":`#569CD6`},children:`false`}),_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})]}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:` }`})}),`\n`,_jsx(_components.span,{className:`line`,children:_jsx(_components.span,{style:{color:`#1F2328`,\"--shiki-dark\":`#D4D4D4`},children:`}`})}),`\n`]})})}),`\n`,_jsxs(_components.p,{children:[`For the bill-doubled case, the key (or model, if the agent groups by `,_jsx(_components.code,{children:`model`}),` instead) whose weekly line jumped is your culprit. For verification, ask the agent to re-run the same query after a fix ships: a successful repoint shows the key’s weekly `,_jsx(_components.code,{children:`total_usage`}),` falling off a cliff at the deploy date, like the sample above.`]}),`\n`,_jsxs(_components.p,{children:[`Filter values must match what the dimension stores internally, so agents should filter on `,_jsx(_components.code,{children:`model`}),` slugs (which they know exactly) and group by `,_jsx(_components.code,{children:`api_key_id`}),` rather than filtering on resolved key names. Save the prompts that worked and re-run them monthly; the copy-prompt at the top chains all of these recipes into one full review.`]}),`\n`,_jsx(Heading,{level:`2`,id:`next-steps`,children:`Next steps`}),`\n`,_jsxs(_components.ul,{children:[`\n`,_jsxs(_components.li,{children:[`Read the `,_jsx(_components.a,{href:`/api/api-reference/analytics/query-analytics-data`,children:`Analytics API reference`}),` for exact request and response schemas.`]}),`\n`,_jsxs(_components.li,{children:[`Drill from an aggregate into individual requests with the `,_jsx(_components.code,{children:`generation_id`}),` dimension, then inspect them with the `,_jsx(_components.a,{href:`https://github.com/OpenRouterTeam/skills/tree/main/skills/openrouter-generations`,children:`openrouter-generations skill`}),`.`]}),`\n`,_jsxs(_components.li,{children:[`Set `,_jsx(_components.a,{href:`/api/api-reference/api-keys/update-an-api-key`,children:`credit limits on keys`}),` once you know which ones drift.`]}),`\n`,_jsxs(_components.li,{children:[`Add `,_jsx(_components.a,{href:`/cookbook/administration/usage-accounting`,children:`usage accounting`}),` to get per-request cost in your own logs.`]}),`\n`,_jsxs(_components.li,{children:[`Use `,_jsx(_components.a,{href:`/guides/best-practices/prompt-caching`,children:`prompt caching`}),` where this review showed low cache rates on prompt-heavy traffic.`]}),`\n`]})]})}function MDXContent(props={}){const{wrapper:MDXLayout}={..._provideComponents(),...props.components};return MDXLayout?_jsx(MDXLayout,{...props,children:_jsx(_createMdxContent,{...props})}):_createMdxContent(props)}return{CopyPromptButton,default:MDXContent};function _missingMdxReference(id,component){throw new Error(`Expected `+(component?`component`:`object`)+\" `\"+id+\"` to be defined: you likely forgot to import, pass, or provide it.\")}","frontmatter":{},"scope":{"config":{},"pageMetadata":{"publish":true,"title":"Control Costs with the Analytics API","description":"Hand your coding agent a management key and the analytics skill, then ask it where your money is going","seoTitle":"Control Costs with the Analytics API | OpenRouter Cookbook","og:site_name":"OpenRouter Documentation","og:title":"Control Costs with the OpenRouter Analytics API","og:description":"Use your coding agent, the Analytics API, and the openrouter-analytics skill to find wasted spend, drill into the keys causing it, and verify the fix.","og:image":"https://openrouter.ai/dynamic-og?title=Control%20Costs%20with%20Analytics\u0026description=Find%20wasted%20spend%20with%20your%20coding%20agent","og:image:width":1200,"og:image:height":630,"twitter:card":"summary_large_image","twitter:site":"@OpenRouter","noindex":false,"nofollow":false,"icon":"chart-line","href":"/cookbook/administration/analytics-cost-control"}}},"mdxSourceWithNoJs":null,"noJsSplices":[[152,728,")=>null;function _cr"]]},"mdxExtracts":{"tableOfContents":[{"title":"Before you start","slug":"before-you-start","depth":2,"children":[]},{"title":"What you’re building","slug":"what-you’re-building","depth":2,"children":[]},{"title":"Setup: install the skill and discover the schema","slug":"setup-install-the-skill-and-discover-the-schema","depth":2,"children":[]},{"title":"Recipe: which models burn the most?","slug":"recipe-which-models-burn-the-most","depth":2,"children":[]},{"title":"Recipe: which keys drive a model’s cost?","slug":"recipe-which-keys-drive-a-model’s-cost","depth":2,"children":[]},{"title":"Recipe: what did the money actually buy?","slug":"recipe-what-did-the-money-actually-buy","depth":2,"children":[{"title":"Separate batch traffic","slug":"separate-batch-traffic","depth":3,"children":[]}]},{"title":"Recipe: what did each token kind cost?","slug":"recipe-what-did-each-token-kind-cost","depth":2,"children":[]},{"title":"Recipe: is my workload prompt-heavy?","slug":"recipe-is-my-workload-prompt-heavy","depth":2,"children":[]},{"title":"Recipe: did spend actually change?","slug":"recipe-did-spend-actually-change","depth":2,"children":[]},{"title":"Next steps","slug":"next-steps","depth":2,"children":[]}],"codeExamples":{}},"jsLinks":["/docs/_mintlify/custom/openrouter-d02e98a0/c1fa764d8c861cbd/anchor-scroll-fix.js","/docs/_mintlify/custom/openrouter-d02e98a0/c43dc2bfe5fa3c92/auth-cta.js"],"dynamicCss":".mint-mt-3 {\n margin-top: 0.75rem\n}\n.mint-inline-flex {\n display: inline-flex\n}\n.mint-cursor-pointer {\n cursor: pointer\n}\n.mint-items-center {\n align-items: center\n}\n.mint-gap-2 {\n gap: 0.5rem\n}\n.mint-rounded-lg {\n border-radius: 0.5rem\n}\n.mint-border {\n border-width: 1px\n}\n.mint-border-green-200 {\n --tw-border-opacity: 1;\n border-color: rgb(187 247 208 / var(--tw-border-opacity, 1))\n}\n.mint-px-3 {\n padding-left: 0.75rem;\n padding-right: 0.75rem\n}\n.mint-py-2 {\n padding-top: 0.5rem;\n padding-bottom: 0.5rem\n}\n.mint-font-semibold {\n font-weight: 600\n}\n.dark\\:mint-border-green-900:is(.dark *) {\n --tw-border-opacity: 1;\n border-color: rgb(20 83 45 / var(--tw-border-opacity, 1))\n}","subdomain":"openrouter-d02e98a0","siteName":"OpenRouter | Documentation","title":"Control Costs with the OpenRouter Analytics API","contextualStarterQuestions":[],"navMaps":{"firstHrefInVersion":{},"firstHrefInLanguage":{}},"lastModified":"2026-11-18T16:25:48.558Z","originalFileLocation":"cookbook/administration/analytics-cost-control.mdx"},"headHtml":"\u003cscript data-mint-head>(function(){var d=document.documentElement,p=null;if(!p){try{p=localStorage.getItem(\"isDarkMode\")}catch(e){}if(p==='true')p='dark';else if(p==='false')p='light';if(p!=='light'\u0026\u0026p!=='dark'\u0026\u0026p!=='system')p=\"dark\"}var r=p==='system'?(matchMedia(\"(prefers-color-scheme: dark)\").matches?'dark':'light'):p;d.classList.remove(r==='dark'?'light':'dark');d.classList.add(r);d.setAttribute(\"data-theme-preference\",p)})();\u003c/script>\u003cmeta data-mint-head name=\"description\" content=\"Hand your coding agent a management key and the analytics skill, then ask it where your money is going\">\u003cmeta data-mint-head name=\"application-name\" content=\"OpenRouter | Documentation\">\u003cmeta data-mint-head name=\"generator\" content=\"Mintlify\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"192x192\" href=\"/docs/mintlify-assets/_mintlify/favicons/openrouter-d02e98a0/EmhfYwFO2tvnkuV-/_generated/favicon/android-chrome-192x192.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"apple-touch-icon\" type=\"image/png\" sizes=\"180x180\" href=\"/docs/mintlify-assets/_mintlify/favicons/openrouter-d02e98a0/EmhfYwFO2tvnkuV-/_generated/favicon/apple-touch-icon.png\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"16x16\" href=\"/docs/mintlify-assets/_mintlify/favicons/openrouter-d02e98a0/EmhfYwFO2tvnkuV-/_generated/favicon/favicon-16x16.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"32x32\" href=\"/docs/mintlify-assets/_mintlify/favicons/openrouter-d02e98a0/EmhfYwFO2tvnkuV-/_generated/favicon/favicon-32x32.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"shortcut icon\" type=\"image/x-icon\" href=\"/docs/mintlify-assets/_mintlify/favicons/openrouter-d02e98a0/EmhfYwFO2tvnkuV-/_generated/favicon/favicon.ico\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"192x192\" href=\"/docs/mintlify-assets/_mintlify/favicons/openrouter-d02e98a0/EmhfYwFO2tvnkuV-/_generated/favicon-dark/android-chrome-192x192.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"16x16\" href=\"/docs/mintlify-assets/_mintlify/favicons/openrouter-d02e98a0/EmhfYwFO2tvnkuV-/_generated/favicon-dark/favicon-16x16.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"32x32\" href=\"/docs/mintlify-assets/_mintlify/favicons/openrouter-d02e98a0/EmhfYwFO2tvnkuV-/_generated/favicon-dark/favicon-32x32.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"shortcut icon\" type=\"image/x-icon\" href=\"/docs/mintlify-assets/_mintlify/favicons/openrouter-d02e98a0/EmhfYwFO2tvnkuV-/_generated/favicon-dark/favicon.ico\" media=\"(prefers-color-scheme: dark)\">\u003cmeta data-mint-head property=\"og:title\" content=\"Control Costs with the OpenRouter Analytics API\">\u003cmeta data-mint-head property=\"og:description\" content=\"Use your coding agent, the Analytics API, and the openrouter-analytics skill to find wasted spend, drill into the keys causing it, and verify the fix.\">\u003cmeta data-mint-head property=\"og:site_name\" content=\"OpenRouter Documentation\">\u003cmeta data-mint-head property=\"og:url\" content=\"https://openrouter.ai/docs/cookbook/administration/analytics-cost-control\">\u003cmeta data-mint-head property=\"og:type\" content=\"website\">\u003cmeta data-mint-head property=\"og:image\" content=\"https://openrouter.ai/dynamic-og?title=Control%20Costs%20with%20Analytics\u0026amp;description=Find%20wasted%20spend%20with%20your%20coding%20agent\">\u003cmeta data-mint-head property=\"og:image:width\" content=\"1200\">\u003cmeta data-mint-head property=\"og:image:height\" content=\"630\">\u003cmeta data-mint-head name=\"twitter:card\" content=\"summary_large_image\">\u003cmeta data-mint-head name=\"twitter:title\" content=\"Control Costs with the OpenRouter Analytics API\">\u003cmeta data-mint-head name=\"twitter:description\" content=\"Use your coding agent, the Analytics API, and the openrouter-analytics skill to find wasted spend, drill into the keys causing it, and verify the fix.\">\u003cmeta data-mint-head name=\"twitter:image\" content=\"https://openrouter.ai/dynamic-og?title=Control%20Costs%20with%20Analytics\u0026amp;description=Find%20wasted%20spend%20with%20your%20coding%20agent\">\u003cmeta data-mint-head name=\"twitter:image:width\" content=\"1200\">\u003cmeta data-mint-head name=\"twitter:image:height\" content=\"630\">\u003cmeta data-mint-head name=\"msapplication-config\" content=\"/docs/mintlify-assets/_mintlify/favicons/openrouter-d02e98a0/EmhfYwFO2tvnkuV-/_generated/favicon/browserconfig.xml\">\u003cmeta data-mint-head name=\"apple-mobile-web-app-title\" content=\"OpenRouter | Documentation\">\u003cmeta data-mint-head name=\"msapplication-TileColor\" content=\"#7624F4\">\u003clink data-mint-head rel=\"canonical\" href=\"https://openrouter.ai/docs/cookbook/administration/analytics-cost-control\">\u003clink data-mint-head rel=\"alternate\" type=\"application/xml\" href=\"/docs/sitemap.xml\">\u003clink data-mint-head rel=\"alternate\" type=\"text/markdown\" href=\"/docs/cookbook/administration/analytics-cost-control.md\">\u003cscript data-mint-head type=\"application/ld+json\">{\"@context\":\"https://schema.org\",\"@graph\":[{\"@type\":\"Organization\",\"@id\":\"https://openrouter.ai/#organization\",\"name\":\"OpenRouter | Documentation\",\"url\":\"https://openrouter.ai\",\"logo\":{\"@type\":\"ImageObject\",\"url\":\"https://mintcdn.com/openrouter-d02e98a0/ksNSeB_K7gD-BUDh/assets/logo-v2-light.svg?fit=max\u0026auto=format\u0026n=ksNSeB_K7gD-BUDh\u0026q=85\u0026s=690e45873843519efe02bcbe45795dac\"}},{\"@type\":\"WebSite\",\"@id\":\"https://openrouter.ai/docs#website\",\"name\":\"OpenRouter | Documentation\",\"url\":\"https://openrouter.ai/docs\",\"publisher\":{\"@id\":\"https://openrouter.ai/#organization\"}},{\"@type\":\"WebPage\",\"@id\":\"https://openrouter.ai/docs/cookbook/administration/analytics-cost-control#webpage\",\"url\":\"https://openrouter.ai/docs/cookbook/administration/analytics-cost-control\",\"name\":\"Control Costs with the Analytics API\",\"description\":\"Hand your coding agent a management key and the analytics skill, then ask it where your money is going\",\"dateModified\":\"2026-11-18T16:25:48.558Z\",\"isPartOf\":{\"@id\":\"https://openrouter.ai/docs#website\"},\"breadcrumb\":{\"@id\":\"https://openrouter.ai/docs/cookbook/administration/analytics-cost-control#breadcrumb\"}},{\"@type\":\"BreadcrumbList\",\"@id\":\"https://openrouter.ai/docs/cookbook/administration/analytics-cost-control#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Administration\",\"item\":\"https://openrouter.ai/docs/cookbook/administration/activity-export\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Control Costs with the Analytics API\",\"item\":\"https://openrouter.ai/docs/cookbook/administration/analytics-cost-control\"}]},{\"@type\":[\"Article\",\"TechArticle\"],\"@id\":\"https://openrouter.ai/docs/cookbook/administration/analytics-cost-control#article\",\"headline\":\"Control Costs with the Analytics API\",\"name\":\"Control Costs with the Analytics API\",\"description\":\"Hand your coding agent a management key and the analytics skill, then ask it where your money is going\",\"url\":\"https://openrouter.ai/docs/cookbook/administration/analytics-cost-control\",\"mainEntityOfPage\":{\"@id\":\"https://openrouter.ai/docs/cookbook/administration/analytics-cost-control#webpage\"},\"image\":\"https://openrouter.ai/dynamic-og?title=Control%20Costs%20with%20Analytics\u0026description=Find%20wasted%20spend%20with%20your%20coding%20agent\",\"dateModified\":\"2026-11-18T16:25:48.558Z\",\"publisher\":{\"@id\":\"https://openrouter.ai/#organization\"},\"isPartOf\":{\"@id\":\"https://openrouter.ai/docs#website\"}}]}\u003c/script>","agentPointerHref":"/docs/cookbook/administration/analytics-cost-control.md","siteName":"OpenRouter | Documentation","nav":{"global":null,"tabs":[{"tab":"Docs","icon":"book","pages":[{"title":"Docs","href":"/quickstart"}]},{"tab":"API Reference","icon":"code","pages":[{"title":"API Reference","href":"/api_reference/overview"}]},{"tab":"Client SDKs","icon":"cube","pages":[{"title":"Client SDKs","href":"/client-sdks/overview"}]},{"tab":"Agent SDK","icon":"cube","pages":[{"title":"Agent SDK","href":"/agent-sdk/overview"}]},{"tab":"Cookbook","icon":"book-open","groups":[{"group":"Get Started","icon":"rocket","pages":[{"title":"Quickstart: Build a Chat App","icon":"message","href":"/cookbook/get-started/quickstart"},{"title":"Migrate to OpenRouter in One Prompt","icon":"arrow-right-arrow-left","href":"/cookbook/get-started/migrate-to-openrouter"},{"title":"Enterprise Onboarding Journey","icon":"route","href":"/cookbook/get-started/enterprise-onboarding-journey"},{"title":"Enterprise Quickstart","icon":"building","href":"/cookbook/get-started/enterprise-quickstart"},{"title":"Free Models Router","icon":"gift","href":"/cookbook/get-started/free-models-router-playground"}]},{"group":"Working with Coding Agents","icon":"robot","pages":[{"title":"Auto-Approve Coding Agent Permission Prompts with Jev","icon":"shield-check","href":"/cookbook/coding-agents/auto-approve-permission-prompts-with-jev"},{"title":"Automatic Code Review","icon":"magnifying-glass-chart","href":"/cookbook/coding-agents/automatic-code-review"},{"title":"Claude Code","icon":"terminal","href":"/cookbook/coding-agents/claude-code-integration"},{"title":"Claude Desktop","icon":"desktop","href":"/cookbook/coding-agents/claude-desktop-integration"},{"title":"Codex CLI","icon":"code","href":"/cookbook/coding-agents/codex-cli"},{"title":"Codex Desktop App","icon":"desktop","href":"/cookbook/coding-agents/codex-desktop"},{"title":"Cursor","icon":"desktop","href":"/cookbook/coding-agents/cursor-integration"},{"title":"Deploy Coding Agents Across Your Company without handing out API keys","icon":"shield-check","href":"/cookbook/coding-agents/deploy-coding-agents"},{"title":"Hermes Agent","icon":"bolt","href":"/cookbook/coding-agents/hermes-integration"},{"title":"Junie CLI","icon":"terminal","href":"/cookbook/coding-agents/junie"},{"title":"Using MCP Servers with OpenRouter","icon":"server","href":"/cookbook/coding-agents/mcp-servers"},{"title":"OpenClaw 🦞","icon":"messages","href":"/cookbook/coding-agents/openclaw-integration"},{"title":"OpenCode","icon":"terminal","href":"/cookbook/coding-agents/opencode-integration"}]},{"group":"Building Agents","icon":"hammer","pages":[{"title":"Add Human-in-the-Loop Controls to an Agent SDK Agent","icon":"user-check","href":"/cookbook/building-agents/hitl-tools"},{"title":"Gate Agent Tool Calls with Jev","icon":"shield-check","href":"/cookbook/building-agents/gate-tool-calls-with-jev"},{"title":"Build a Long-Horizon Agent","icon":"clock","href":"/cookbook/building-agents/long-horizon-agents"},{"title":"Build a Token-Efficient Review Agent","icon":"lightbulb","href":"/cookbook/building-agents/advisor-server-tool"},{"title":"Delegate Routine Work to Cheaper Models","icon":"sitemap","href":"/cookbook/building-agents/subagent-server-tool"},{"title":"Build Your Own Agent TUI","icon":"hammer","href":"/cookbook/building-agents/create-agent-harness-tui"},{"title":"Build Your Own Headless Agent","icon":"robot","href":"/cookbook/building-agents/create-headless-agent"},{"title":"Compact Long Agent Sessions with ScaleDown Summarize","icon":"compress","href":"/cookbook/building-agents/compact-agent-sessions-with-scaledown"},{"title":"Detect Stuck and Looping Agents with ScaleDown Classify","icon":"rotate-exclamation","href":"/cookbook/building-agents/detect-agent-loops-with-scaledown"}]},{"group":"Video Generation","icon":"film","pages":[{"title":"Choose a Video Generation Model","icon":"filter","href":"/cookbook/video-generation/choose-video-model"},{"title":"Generate and Download a Video from Text","icon":"wand-magic-sparkles","href":"/cookbook/video-generation/text-to-video"},{"title":"Get Video Results with Webhooks","icon":"link","href":"/cookbook/video-generation/video-generation-webhooks"},{"title":"Guide a Video with Reference Images","icon":"images","href":"/cookbook/video-generation/reference-to-video"},{"title":"Turn an Image into a Video","icon":"image","href":"/cookbook/video-generation/image-to-video"},{"title":"Use Provider-Specific Video Options","icon":"sliders","href":"/cookbook/video-generation/provider-specific-video-options"}]},{"group":"Audio","icon":"microphone","pages":[{"title":"Turn Any Text into a Two-Voice Podcast","sidebarTitle":"Two-Voice Podcast","icon":"podcast","href":"/cookbook/audio/two-voice-podcast"},{"title":"Add a Push-to-Talk Voice Button to Your App","sidebarTitle":"Push-to-Talk Button","icon":"microphone-lines","href":"/cookbook/audio/push-to-talk-button"}]},{"group":"Image Understanding","icon":"eye","pages":[{"title":"Find and Mark Objects in Any Image","sidebarTitle":"Find Objects in Images","icon":"crosshairs","href":"/cookbook/image-understanding/find-objects-in-images"}]},{"group":"Image Generation","icon":"image","pages":[{"title":"Enhance Image Generation with Presets","icon":"wand-magic-sparkles","href":"/cookbook/image-generation/preset-enhanced-images"}]},{"group":"Evaluate \u0026 Optimize","icon":"flask","pages":[{"title":"Distillation","icon":"filter","href":"/cookbook/evaluate-and-optimize/distillation"},{"group":"Model Migrations","icon":"arrow-right-arrow-left","pages":[{"title":"Claude 5.5 Migration Guide","icon":"arrow-right-arrow-left","href":"/cookbook/evaluate-and-optimize/model-migrations/claude-5-5"},{"title":"Claude Fable 5.1 Migration Guide","icon":"arrow-right-arrow-left","href":"/cookbook/evaluate-and-optimize/model-migrations/fable-5-1"},{"title":"Claude Opus 5 Migration Guide","icon":"arrow-right-arrow-left","href":"/cookbook/evaluate-and-optimize/model-migrations/opus-5"},{"title":"Claude 5 Sonnet Migration Guide","icon":"arrow-right-arrow-left","href":"/cookbook/evaluate-and-optimize/model-migrations/sonnet-5"},{"title":"Claude 4.7 Migration Guide","icon":"arrow-right-arrow-left","href":"/cookbook/evaluate-and-optimize/model-migrations/claude-4-7"},{"title":"Claude 4.6 Migration Guide","icon":"arrow-right-arrow-left","href":"/cookbook/evaluate-and-optimize/model-migrations/claude-4-6"},{"title":"Gemini 4 Migration Guide","icon":"arrow-right-arrow-left","href":"/cookbook/evaluate-and-optimize/model-migrations/gemini-4"},{"title":"GPT-6 Migration Guide","icon":"arrow-right-arrow-left","href":"/cookbook/evaluate-and-optimize/model-migrations/gpt-6"},{"title":"GPT-5.6 Migration Guide","icon":"arrow-right-arrow-left","href":"/cookbook/evaluate-and-optimize/model-migrations/gpt-5-6"},{"title":"GPT-5.4 Migration Guide","icon":"arrow-right-arrow-left","href":"/cookbook/evaluate-and-optimize/model-migrations/gpt-5-4"}]},{"title":"RAG with Embeddings \u0026 Rerank","icon":"magnifying-glass-chart","href":"/cookbook/evaluate-and-optimize/rag"},{"title":"Cut LLM Cost with a Jev-Verified Cascade","icon":"shield-check","href":"/cookbook/evaluate-and-optimize/jev-verified-cascade"},{"title":"Classify Reddit and YouTube Comments with Jev and ScrapeCreators","icon":"comments","href":"/cookbook/evaluate-and-optimize/classify-reddit-comments-with-jev"},{"title":"Classify and Tag Text at Scale with Jev","icon":"tags","href":"/cookbook/evaluate-and-optimize/jev-classification"},{"title":"Extract Invoice Fields with ScaleDown Extract","icon":"file-invoice-dollar","href":"/cookbook/evaluate-and-optimize/extract-invoice-fields-with-scaledown"},{"title":"Red Teaming","icon":"flask-vial","href":"/cookbook/evaluate-and-optimize/red-teaming"}]},{"group":"Administration","icon":"chart-line","pages":[{"title":"Activity Export","icon":"file-export","href":"/cookbook/administration/activity-export"},{"title":"API Key Audit","icon":"magnifying-glass","href":"/cookbook/administration/api-key-audit"},{"title":"API Key Rotation","description":"Securely rotate your OpenRouter API keys","icon":"rotate","href":"/cookbook/administration/api-key-rotation"},{"title":"Control Costs with the Analytics API","icon":"chart-line","href":"/cookbook/administration/analytics-cost-control"},{"title":"Crypto API","description":"Coinbase Commerce API deprecation","icon":"bitcoin","href":"/cookbook/administration/crypto-api"},{"title":"Data API","icon":"database","href":"/cookbook/administration/data-api"},{"title":"Organization Management","icon":"building-user","href":"/cookbook/administration/organization-management"},{"title":"Adding a Tax ID to Your Invoices","icon":"file-invoice","href":"/cookbook/administration/tax-id"},{"title":"Usage Accounting","icon":"calculator","href":"/cookbook/administration/usage-accounting"},{"title":"User Tracking","icon":"users","href":"/cookbook/administration/user-tracking"}]}]}]},"basePath":"","title":"Control Costs with the OpenRouter Analytics API","deploymentId":"dpl_J2QdBVuN65x4CCtpqPb371YPt4xi","buildId":"dpl_J2QdBVuN65x4CCtpqPb371YPt4xi","dynamics":[],"chunks":[]}