57%
of orgs have no observability
v0 · Agent observability
AI agents fail silently and bill loudly.
Most developers find out their agent broke from an angry user. Farol tells you first.
The gap
You can’t fix what you never see
40×
cost spikes go undetected
3
lines of code to get started
sdk.ts
import { farol } from '@farol/sdk'
farol.init({ apiKey: process.env.FAROL_KEY }) // that’s itProduct
Built for builders who ship agents, not slides
Live run dashboard
Watch sessions as they happen: status, latency, and tool calls in one stream—no digging through logs.
Cost anomaly detection
Token spend and provider charges tracked per run. Get alerted when usage jumps before finance does.
Full trace inspector
Open any run: prompts, retrievals, model hops, and errors—reconstructed end to end for fast debugging.
Compare
Farol vs. the usual suspects
| Capability | Farol | Braintrust | LangSmith |
|---|---|---|---|
| Indie & small-team pricing | Yes | Team tiers | Org-focused |
| Live run stream | Yes | Yes | Yes |
| Cost anomaly alerts | First-class | Evals-first | Usage views |
| Minimal SDK surface | 3 lines | SDK + setup | Heavier config |
| Opinionated for solo devs | Yes | — | — |
Be the first to know when something breaks
Join the waitlist for early access, pricing for small teams, and the public beta timeline.