# Foglamp

> The missing observability layer for AI agents. Costs, latency, tokens, distributed traces, evals, and alerts for every model call - in two lines.

## Docs

- [Instrument with an AI agent](https://docs.foglamp.dev/ai-instrument.md): Instructions for a coding agent to add Foglamp tracing to an app.
- [Health check](https://docs.foglamp.dev/api-reference/ingest/health-check.md): Liveness probe. Returns the current size of the in-memory write buffer.
- [Ingest a batch of traces](https://docs.foglamp.dev/api-reference/ingest/ingest-a-batch-of-traces.md): Validates the payload, resolves the API key, prices each span, and buffers the rows for a flush to ClickHouse. Fire-and-forget: a 202 means the batch was accepted, not yet durably written.
- [API reference](https://docs.foglamp.dev/api-reference/introduction.md): The Foglamp ingest API.
- [Changelog](https://docs.foglamp.dev/changelog.md): What's new in Foglamp.
- [Data model](https://docs.foglamp.dev/concepts/data-model.md): Traces, spans, workflows, runs, agents, and sessions.
- [Projects, keys & billing](https://docs.foglamp.dev/dashboard/account.md): API keys, team roles, plans, and usage quotas.
- [Agents](https://docs.foglamp.dev/dashboard/agents.md): Per-agent rollups — cost, latency, errors — across every run.
- [Alerts](https://docs.foglamp.dev/dashboard/alerts.md): Threshold rules on cost, latency, errors, and eval scores.
- [Cost & pricing](https://docs.foglamp.dev/dashboard/cost.md): How Foglamp prices spans.
- [Evals](https://docs.foglamp.dev/dashboard/evals.md): Score production traffic with code checks or LLM judges.
- [Overview](https://docs.foglamp.dev/dashboard/overview.md): The dashboard home — KPIs, trends, and per-model/agent/workflow breakdowns.
- [Sessions](https://docs.foglamp.dev/dashboard/sessions.md): Follow one conversation or user thread end to end.
- [Traces](https://docs.foglamp.dev/dashboard/traces.md): Inspect a single run — waterfall, span detail, scores, and replay.
- [Workflows](https://docs.foglamp.dev/dashboard/workflows.md): Group multi-call processes into runs and inspect them as a flow.
- [Introduction](https://docs.foglamp.dev/introduction.md): The missing observability layer for the Vercel AI SDK.
- [Quickstart](https://docs.foglamp.dev/quickstart.md): Instrument a Vercel AI SDK app and see your first trace.
- [Configuration](https://docs.foglamp.dev/sdk/configuration.md): Every option accepted by foglamp().
- [SDK overview](https://docs.foglamp.dev/sdk/overview.md): The foglamp() collector and its integration methods.
- [Runtimes & flushing](https://docs.foglamp.dev/sdk/runtimes.md): Make sure spans leave the process in every runtime.
- [AI SDK v4–v6 (wrap)](https://docs.foglamp.dev/sdk/wrap.md): Instrument older AI SDK versions by wrapping the module.
- [Configuration](https://docs.foglamp.dev/self-hosting/configuration.md): Environment variable reference for a self-hosted deployment.
- [Self-hosting](https://docs.foglamp.dev/self-hosting/overview.md): Run the full Foglamp stack on your own infrastructure.
- [Troubleshooting](https://docs.foglamp.dev/troubleshooting.md): Why you might not see traces, and how to fix it.

## OpenAPI Specs

- [openapi](https://docs.foglamp.dev/api-reference/openapi.json)