Logfire
Production-grade observability built on OpenTelemetry for AI applications, agents, APIs, and databases. Includes distributed tracing, evals, and cost tracking with SDKs for Python, JavaScript/TypeScript, and Rust.
Pydantic Logfire is a production-grade observability platform designed specifically for AI applications, agents, APIs, and databases. Built on OpenTelemetry, it provides comprehensive monitoring across your entire stack, not just LLM calls.
Key Features
- Whole-App Tracing: Trace model calls, agent steps, API requests, and database queries in a single timeline
- Multi-Language Support: SDKs for Python, JavaScript/TypeScript, and Rust with OpenTelemetry compatibility for other languages
- Evals & Scoring: Built-in evaluation framework that runs alongside your application logic
- Cost Tracking: Monitor and optimize AI application costs in real-time
- PII Scrubbing: Automatic sensitive data redaction in SDKs
- Postgres-Flavored SQL: Query spans directly using familiar SQL syntax
- Live Debugging: Real-time monitoring of production services with error rates, latency, and traces
- Self-Hosting Option: Deployable on your own infrastructure
Tech / Self-Hosting
- OpenTelemetry Foundation: Built on industry-standard observability protocols
- Self-Hosted Deployment: Available for on-premises or private cloud deployments
- Compliance: SOC 2 Type II, HIPAA, and EU data residency compliant
- Integration: Seamless integration with existing monitoring and logging systems