Fireworks AI usage
Serverless tokens per model, key and day, and the bill's rated line items per day. No prompt, no response, no request.
Access
- Kind
- A read-only key you paste
- What to give
- API key and account id
- Why that one
- A Fireworks API key on the account. Valistry reads the account's usage and billing summary, never a request.
What it attributes on its own
- From the source
- day · model · API key · prompt, completion and cached tokens · what Fireworks rated the day at, per model
- Needs a rule or a trace
- team, workflow and agent unless a rule maps the key · per-request detail — Fireworks reports daily totals · outcome
Connecting it
- Create an API key. Settings → API keys → Create. A key on an administrator's account reads the billing summary; another key reads tokens only. Open ↗
- Copy the account id. The slug after /accounts/ in the console URL.
- Paste both and click Test and connect. Valistry reads the last seven days, reports what it found, then fills thirty.
Fireworks usage read through a gateway or OpenTelemetry carries no bill. This source carries the bill, so where both are connected, connect this one and disconnect the other for Fireworks traffic.