foglamp.dev

mx

mx1.improvmx.com

mx2.improvmx.com

ns

ns2.vercel-dns.com

ns1.vercel-dns.com

caa

pki.goog

letsencrypt.org

sectigo.com

txt

google-site-verification=2AHJNaUy-N9Sxak1euF9PhQ6hozQpIM7KK6vkzdBTQE

v=spf1 include:spf.improvmx.com ~all

dmarc

v=DMARC1; p=none;

User-Agent: *
Allow: /
Disallow: /overview
Disallow: /agents
Disallow: /alerts
Disallow: /evals
Disallow: /platform
Disallow: /sessions
Disallow: /settings
Disallow: /traces
Disallow: /workflows
Disallow: /login
Disallow: /device
Disallow: /reset-password
Disallow: /accept-invitation/

Sitemap: https://foglamp.dev/sitemap.xml

# Foglamp

> The missing observability layer for the Vercel AI SDK. Two lines of code and an
> API key give you costs, latency, token usage, distributed traces, evals, and
> alerts for every generateText / streamText call — across every model your AI
> agents use. Open source (Apache 2.0) and self-hostable.

## Get started
- [Instrument your app (written for coding agents)](https://docs.foglamp.dev/ai-instrument.md): version-aware setup for Vercel AI SDK v4–v7, with agent / workflow / session mapping rules and verification steps.
- [Quickstart](https://docs.foglamp.dev/quickstart): install the SDK, add an API key, and see your first trace.
- [Full docs index for LLMs](https://docs.foglamp.dev/llms.txt): the complete machine-readable documentation tree.

## Product
- [Overview](https://foglamp.dev): what Foglamp is and who it's for.
- [How it works](https://foglamp.dev/how-it-works): an illustrated guide to traces, agents, workflows, sessions, costs, evaluations, and alerts.
- [Pricing](https://foglamp.dev/pricing): usage-based plans — Free ($0), Pro ($49/mo), and Enterprise (custom).
- [Distributed traces](https://foglamp.dev/features/distributed-traces): waterfall every run down to the token, with the exact prompt and response per span.
- [Cost intelligence](https://foglamp.dev/features/cost-intelligence): spend broken down by model, agent, and customer.
- [Evals](https://foglamp.dev/features/evals): score production traffic with code checks and LLM judges.
- [Agents](https://foglamp.dev/features/agents): per-agent spans, latency, and spend with the full call flow.
- [Alerts](https://foglamp.dev/features/alerts): threshold rules on cost, latency, and error rate.
- [SDK](https://foglamp.dev/features/sdk): two lines instrument every generateText / streamText call.

## Foglamp developer resources
- [Foglamp API reference](https://docs.foglamp.dev/api-reference/introduction): the ingest API, authentication, and error shapes.
- [Foglamp OpenAPI spec](https://docs.foglamp.dev/api-reference/openapi.json): machine-readable API specification (also at https://foglamp.dev/openapi.json).
- [Foglamp changelog](https://docs.foglamp.dev/changelog): what shipped, release by release.
- [Foglamp source code](https://github.com/foglamp-labs/foglamp): GitHub repository (Apache 2.0, self-hostable).

## Notes for agents
- Every marketing page on https://foglamp.dev is also served as markdown: request it
  with `Accept: text/markdown`, or append `.md` to the path (e.g.
  https://foglamp.dev/pricing.md).
- Nonexistent paths return a real 404 with a markdown body pointing back here.