Pydantic Logfire

What it does

Pydantic Logfire takes in OpenTelemetry traces, logs, and metrics from AI applications, including LLM calls, tool calls, database queries, and API requests, and stores them as queryable data. It produces trace views, eval results, dashboards, alerts, and cost tracking, all reachable through standard SQL. Native SDKs cover Python, TypeScript, and Rust, and any language with an OpenTelemetry SDK can send data.

People use Logfire to find why an AI run failed, score outputs against datasets, and change prompts or model settings without redeploying through managed variables. The coding-agent MCP server lets an assistant such as Claude Code or Cursor query production telemetry directly. The AI Gateway routes model calls through one endpoint with spend caps per key, person, and project.

Logfire suits developers and teams running production AI or services who want OpenTelemetry-based instrumentation rather than a proprietary wrapper. Access starts with a free Personal plan of 10 million records per month, with ingestion paused after that cap. Team starts at $49 per month, and Enterprise offers custom cloud, dedicated, or self-hosted deployment. Data retention is 30 days on Personal and Team, and up to 90 days on Growth.

What it can do

• Trace AI agent runs across LLM calls, tool calls, database queries, and APIs in one view
• Run evals with datasets, code assertions, LLM-as-a-judge rubrics, and human annotations
• Query traces, logs, and metrics with standard PostgreSQL SQL
• Route model calls through the AI Gateway with per-key, per-person, and per-project spend caps
• Change prompts, model settings, and managed variables without redeploying (OpenFeature-compatible)
• Build dashboards and alerts on any queryable signal, including eval scores, latency, and spend

Works with

• OpenAI
• Anthropic
• LangChain
• LangGraph
• LlamaIndex
• CrewAI
• Vercel AI SDK
• Pydantic AI
• OpenTelemetry (standard OTel for any language)

Before you choose

• Personal plan ingestion pauses at the 10 million record monthly cap, and ad-hoc overage payment is not offered on Personal.
• Data retention is 30 days on Personal and Team and up to 90 days on Growth; longer retention requires Enterprise.
• Full AI Gateway access with Data Loss Prevention is an add-on on Enterprise, and built-in provider usage carries a markup on lower plans.

Listed under

• Coding
• DevOps and observability
• Sandboxes and infrastructure
• AI agents
• Developers
• Evals
• LLM monitoring
• Observability
• SQL

Get started

Personal plan has 10 million records per month and three projects; uv add logfire installs the SDK.

1. Create a free account at pydantic.dev/signup.
2. Install the SDK with uv add logfire (or run uvx logfire-cli wizard for guided setup with a coding agent).
3. Call logfire.configure() in your application and emit a test event such as logfire.info('Hello {name}', name='world') to confirm traces arrive.

Every plan includes $20 of free usage per month (10 million records). Team and Growth are billed per million records above that allowance, with optional price caps. Enterprise is custom and invoiced. Personal is free and hard-capped.

Source: pydantic.dev/pricing, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Pydantic Logfire: AI Observability for LLMs, Apps & RAG ↗pydantic.dev
• Pricing and Plans for Pydantic Logfire ↗pydantic.dev
• Pydantic Docs - Validation, AI Agents, Logfire Observability ↗pydantic.dev

Company website

pydantic.dev: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.