logfire Claude Code Plugin
Claude Code PluginClaude CodeMonitoringMonitoring & ObservabilityAI Agents & AI App DevelopmentClaude Code Customization & WorkflowThe explanation below is AI-generated. Please verify it against the sources.
This plugin adds Logfire observability support to Python projects by giving Claude Code (and similarly Codex and Cursor) a set of skills and commands for instrumenting code, debugging via traces, and querying telemetry data. It automatically detects frameworks such as FastAPI, httpx, asyncpg, and SQLAlchemy and wires up the corresponding Logfire instrumentation. It bundles slash commands for instrumenting an app, investigating errors, and running queries, plus a skill for opening Logfire's web views directly from the editor. Data collection and querying are handled through a connected Logfire MCP server, which can be pointed at the US region, EU region, or a self-hosted deployment. Using it requires authenticating to a Logfire account or providing a write token.
About the service
Logfire is an observability platform from Pydantic for monitoring applications and AI/LLM workloads. Its console lets users search and query collected data with SQL, and it offers enterprise options such as self-hosting, SOC2 Type II certification, HIPAA compliance, and an EU data region. Free usage is available without a credit card, alongside paid enterprise plans.
What you can do with logfire
- Automatically detect frameworks (e. g.
- Investigate application errors by pulling in Logfire trace data through /debug
- Run interactive queries against traces, logs, and metrics, or add query code, with /query
- Open Logfire project pages, live trace views, and Explore filters directly from the editor via the logfire-ui skill
- Reference bundled guidance (SKILL. md) covering core Logfire setup patterns such as configuration, instrumentation, structured logging, and AI/LLM tracing
- Query Logfire data through a connected MCP server, configurable for the US region, EU region, or a self-hosted instance