Usage & Cost
Lattis attributes every request and tracks token usage broken down by app, model, project, and conversation. For paid cloud models it also estimates the dollar cost from the provider’s reported token counts.
What’s tracked
Section titled “What’s tracked”- Per-app — which client/tool made the request.
- Per-model — local or cloud model id used.
- Per-project — work attributed to a project context, and per branch within it.
- Per-conversation — spend on each conversation in Lattis’s own chat.
Token counts come straight from the inference engine (local) or the provider’s usage fields (cloud), so the breakdown reflects actual consumption. Within a project branch, spend is further broken down per model, so you can see which model a branch’s work is going to.
Persistence
Section titled “Persistence”Usage is recorded to an on-disk store (telemetry.db in the data directory) and
rehydrated when the daemon restarts, so your totals survive restarts. Costs are
priced at the moment a request is made and stored as recorded, so historical
figures don’t shift when the price table changes. Live-only state — active
connections and the current tokens/second — resets with the daemon.
The web app: spend & activity across machines
Section titled “The web app: spend & activity across machines”Everything above is tracked locally by default — nothing leaves your machine. The Lattis web app at app.getlattis.ai is an optional, hosted dashboard that rolls the same usage and cost data up across your machines and your whole team.
Sign in with GitHub, Google, or a magic link — signing in for the first time creates your account and a personal organization automatically (there is no separate signup). Once you connect your desktop app, you get:
- Overview — spend, cache savings, active users, LLM calls, latency, and error rate at a glance.
- Spend — dollar spend over time, broken down by project and model.
- Traces & Sessions — drill into individual calls and whole sessions, with per-request cost, latency, and tool failures.
- Models — spend and cost-effectiveness per model, so you can decide what runs where.
- Team & Members — invite people, set roles, and see an org-wide leaderboard.
Lattis stays fully local by default; the web app is opt-in, and nothing is sent unless you connect an account.
Connecting your desktop app
Section titled “Connecting your desktop app”In the desktop app, choose Connect account. Your browser opens the web app’s
/desktop/connect handoff; once you’re signed in it hands a one-time code back
to the app over a local loopback callback, and your fleet’s activity begins
flowing into the dashboard. The code only travels in that redirect — it is never
displayed or logged.
The web app is free during our public beta, and free forever for individuals and small teams — see Pricing.
Cost estimation
Section titled “Cost estimation”Cloud providers report token counts, not dollars. Lattis multiplies those counts by a per-model price table (split by billing category — standard input, output, cached-input reads, and cache writes) to produce an estimate. Local models are free, so no cost is shown for them.
Because pricing is an estimate derived from a built-in table, treat the numbers as a close guide rather than an invoice.
Why it matters
Section titled “Why it matters”A single endpoint serving many models and providers makes it easy to lose track of where tokens — and money — go. Per-project attribution answers “what is this project costing?” and per-model breakdown shows whether a task belongs on a local model or a frontier cloud one.
See Cloud Providers to connect a paid provider, and the Control API for the daemon snapshot that surfaces these figures.