nlqdb

Solve · Agent builders

How do I track and query my AI app's token usage and cost per user?

If your LLM app needs to track token usage and cost — per user, per model, per day — log each call as a row and ask in English. nlqdb provisions Postgres from your first goal and runs the GROUP BY in SQL, so 'spend per user this month' is a real query, not arithmetic over a JSON log.

Teams shipping LLM features need to answer 'how much are we spending per customer?' and 'which model cost the most this week?' — but token counts and dollar costs usually land in application logs or a JSON column. Pulling those back to total them in app code (or asking the LLM to add them up) is fragile and doesn't scale; these questions are aggregations, and aggregations want a query planner, not a log scan.

The snippet that solves it.

> total tokens and cost grouped by model this month, highest cost first

What nlqdb does for this

  • Log each call as a typed row — user, model, prompt/completion tokens, cost, timestamp — so cost-per-user and tokens-per-model run as real SQL GROUP BY, not log math.
  • Ask the cost question in English via the `<nlq-data>` element, the `@nlqdb/sdk`, or MCP `nlqdb_query`; every answer returns rows plus the compiled SQL under a trace toggle.
  • Write usage rows with the deterministic `nlqdb_remember` tool or a `POST /v1/run` parameterised INSERT, and report over the same database — no separate analytics store, no ETL.
  • Plans are content-addressed on `(goal-fingerprint, schema-hash)` (`GLOBAL-006`), so a repeated weekly cost rollup hits the cache and returns in single-digit ms.

Drop into any HTML page

<nlq-data goal="total tokens and cost grouped by model this month, highest cost first"></nlq-data>

The first cost question an LLM team asks — spend broken down by model — is one English goal here, not a hand-written GROUP BY over a usage log.

What this replaces

What nlqdb doesn't try to do here

  • No automatic token metering — nlqdb stores and aggregates the usage rows you write; counting tokens and computing cost per call is your app's job (or your provider SDK's).
  • No connecting to your existing logging or billing store — nlqdb provisions and owns the Postgres it queries; bring-your-own-Postgres is roadmap, not shipped.
  • No live streaming cost meter — the table refreshes on query, not via websocket push; realtime dashboards are roadmap, not shipped.

Questions buyers ask

Why not just total token usage in my application code?
Because the questions are aggregations — spend per user, tokens per model, cost per day — and pulling rows back to sum them in app code is fragile and slow as volume grows. Asking the LLM to add them up is worse: arithmetic over a list is a hallucination generator. nlqdb runs the GROUP BY in Postgres and shows you the SQL it ran.
How do the token and cost numbers get into the database?
Write one row per LLM call — user, model, prompt and completion tokens, computed cost, timestamp — with the deterministic `nlqdb_remember` MCP tool or a parameterised INSERT through `POST /v1/run` (`GLOBAL-015`). The row shape stays a trust boundary, built server-side, not LLM-guessed. Then ask the cost questions in English over the same table.
Can I see the SQL behind the cost numbers?
Always — every answer returns the result rows plus the compiled SQL under a trace toggle (`SK-WEB-005`), so you can verify the grain (per call vs per user) before trusting a spend figure. nlqdb never hides the SQL behind the answer.
Is this a replacement for an LLM observability tool like Langfuse or Helicone?
No — those proxy or instrument your calls and capture token and cost automatically, with tracing UIs built for it. nlqdb is the database half: you decide what to log, and you get a SQL query planner over it for ad-hoc 'spend per X' questions without a per-seat dashboard tool. They compose; nlqdb doesn't proxy your traffic.

Where this pain shows up in public

Enduring discussion hubs where you can verify the theme without taking our word for it. We don't quote individual posts; we cite search-result and subreddit URLs that stay live as new threads land.

Try nlqdb in 30 seconds

No sign-in. The anonymous database lasts 72 hours; adopt it with one click if you keep it.

Start with a goal →

Looking at this from a different angle? Browse all solve pages or browse competitor comparisons.