Pydantic Logfire

What is Pydantic Logfire?

One trace across model calls, agents, APIs and database queries.

Who is Pydantic Logfire best for?

Pydantic Logfire is best for developers running AI apps in production who want every step in one trace.

• Teams building AI agents

  Each agent run is traced from prompt to tool calls, with data from Pydantic AI, LangChain, the OpenAI SDK or any OpenTelemetry setup.

• Debugging whole requests

  The model call, the API hop and the database query sit in one trace, so a slow vector search shows up next to the LLM call it delayed.

• Engineers who think in SQL

  Every span can be queried with Postgres-style SQL, including joins and JSON operators, in the app or from a coding agent over MCP.

• Checking prompt and model changes

  Evals are stored on the traces they score and billed as ordinary records, so comparing runs adds no per-score fee.

How much does Pydantic Logfire cost?

Pydantic Logfire is free forever on the Personal plan with 10 million records a month, and its paid plans cost $49 a month for Team or $249 a month for Growth, plus $2 per million records beyond 10 million.

Enterprise is priced on request, as managed cloud, dedicated or self-hosted, with SSO, SCIM, custom retention and SLA-backed support. Extra Team seats are $25 each, up to 12 seats. Team and Growth can set a spending cap, and the first month comes with a money-back guarantee. Non-profits, academics, community groups and pre-funding startups can apply for financial support. The AI Gateway adds a 5% markup on built-in model providers on Personal and Team and 3% on Growth, with no markup when you bring your own keys. Full pricing on pydantic.dev. We checked these prices on 27 Sep 2026.

How do readers rate Pydantic Logfire?

No one has rated Pydantic Logfire yet. Be the first.

What are the best Pydantic Logfire alternatives?

Three tools on Recaply do Pydantic Logfire's job. All three are open source, so you can run them on your own server.

AgentOpsOpen Source

Choose it when your traces are mostly agent runs and their token costs, and you want a tracer you can self-host under MIT.

LaminarOpen Source

Choose it when you want open-source tracing built around agent runs that you can self-host under Apache 2.0.

SigNozOpen Source

Choose it when you need LLM traces next to infrastructure metrics and logs in a tool you can self-host for free

Questions about Pydantic Logfire