What is Pydantic Logfire?

Pydantic is an AI engineering stack for teams building with AI. It combines Pydantic AI for building agents in Python or TypeScript with Pydantic Logfire for tracing production agents, turning failures into eval cases, and improving what ships. Built on OpenTelemetry, it connects model calls and tool use to the APIs, databases, queues, and infrastructure behind each result.

How to use Pydantic Logfire?

Start for free with no credit card required, which includes 10M telemetry records per month. Bring your agents written in Python or TypeScript and onboard with your coding agent. Use Logfire to trace production agents, turn failures into eval cases, and continuously improve your agents with insights from traces and evals. Protect data and control spend with Logfire AI Gateway.

Core features of Pydantic Logfire

Pydantic AI: build agents you'll actually ship to production in Python or TypeScript
Pydantic Logfire: trace production agents and turn failures into eval cases
One trace across agent, model calls, APIs, and database
Built on OpenTelemetry with first-party SDKs for Python, JavaScript/TypeScript, and Rust
Integrations with agent frameworks, web frameworks, databases, and infrastructure
Logfire AI Gateway to protect data and control spend
Infrastructure monitoring for hosts, clusters, containers, and cloud metrics
Dashboards and alerts