Agent frameworks
Agno
Curated Agents view
Logfire integrations
Every model, framework, service and language in one trace.
Native instrumentation where it matters. OpenTelemetry everywhere else.
106catalogued integrations
9parts of your stack
13languages and runtimes
Catalog reviewed Sep 18, 2026
One requestOne trace
- Agent frameworkPydantic AI0 ms
- Model providerAnthropic184 ms
- HTTP clientHTTPX241 ms
- DatabasePostgreSQL319 ms
Correlated by OpenTelemetry context
111 setup paths
All integrations.
Every setup link opens the maintained guide. Where coverage has been verified, agent cards say whether Logfire builds a curated Agents view or keeps the complete trace in Live and Explore.
How integration support works
First-party SDK A Pydantic-maintained SDK sends telemetry directly to Logfire.
Automatic instrumentation A maintained integration instruments the library or framework without hand-building spans.
Curated OpenTelemetry A maintained setup guide connects its OpenTelemetry, with product-view coverage documented framework by framework.
Generic OpenTelemetry Standard OTLP data stays searchable in Live and Explore without a Logfire-specific SDK.
Showing 106 of 106 integrations
Agent frameworks
AutoGen
Curated Agents view
Agent frameworks
Claude Agent SDK
Complete trace in Live & Explore
Agent frameworks
CrewAI
Curated Agents view
Agent frameworks
Eino
Complete trace in Live & Explore
Agent frameworks
Eve
Curated Agents view
Agent frameworks
Genkit
Curated Agents view
Agent frameworks
Google ADK
Curated Agents view
Agent frameworks
Haystack
Complete trace in Live & Explore
Agent frameworks
LangChain
Agent frameworks
LangChain.js
Complete trace in Live & Explore
Agent frameworks
LangGraph
Curated Agents view
Agent frameworks
Letta
Complete trace in Live & Explore
Agent frameworks
LlamaIndex
Agent frameworks
LlamaIndex.TS
Complete trace in Live & Explore
Agent frameworks
Magentic
Agent frameworks
Mastra
Curated Agents view
Agent frameworks
MCP
Agent frameworks
Microsoft Agent Framework
Curated Agents view
Agent frameworks
Mirascope
Agent frameworks
OpenAI Agents SDK
Curated Agents view
Agent frameworks
OpenAI Agents SDK (TS)
Complete trace in Live & Explore
Agent frameworks
Pydantic AI
Curated Agents view
Agent frameworks
Rig
Curated Agents view
Agent frameworks
Semantic Kernel
Curated Agents view
Agent frameworks
Semantic Kernel (.NET)
Curated Agents view
Agent frameworks
smolagents
Curated Agents view
Agent frameworks
Strands Agents
Curated Agents view
Agent frameworks
Vercel AI SDK
Curated Agents view
Agent frameworks
VoltAgent
Curated Agents view
Models and LLM clients
Anthropic
Models and LLM clients
DSPy
Models and LLM clients
Google GenAI
Models and LLM clients
Instructor
Complete trace in Live & Explore
Models and LLM clients
LiteLLM
Models and LLM clients
OpenAI
Models and LLM clients
OpenRouter Broadcast
Web frameworks
AIOHTTP
Web frameworks
ASGI
Web frameworks
Cloudflare Workers
Web frameworks
Deno
Web frameworks
Django
Web frameworks
Express
Web frameworks
FastAPI
Web frameworks
Flask
Web frameworks
Gunicorn
Web frameworks
Litestar
Web frameworks
Next.js
Web frameworks
Starlette
Web frameworks
WSGI
Databases
Asyncpg
Databases
BigQuery
2 support paths
Elasticsearch
2 support paths
MongoDB
2 support paths
MySQL
2 support paths
PostgreSQL
2 support paths
Redis
Databases
SQLAlchemy
Databases
SQLite3
Databases
SurrealDB
HTTP clients
AIOHTTP client
HTTP clients
HTTPX
HTTP clients
Requests
Queues and pipelines
Apache Airflow
Queues and pipelines
Celery
Queues and pipelines
FastStream
Logging
log crate
Logging
Loguru
Logging
opentelemetry
Logging
Logging
Python logging
Logging
Structlog
Logging
tracing
Infrastructure and services
ActiveMQ
Infrastructure and services
Apache Cassandra
Infrastructure and services
Apache Flink
Infrastructure and services
Apache HTTP Server
Infrastructure and services
Apache Kafka
Infrastructure and services
Apache Spark
Infrastructure and services
Apache Tomcat
Infrastructure and services
AWS Lambda
Infrastructure and services
Docker
Infrastructure and services
HAProxy
Infrastructure and services
Kubernetes
Infrastructure and services
Memcached
Infrastructure and services
NGINX
Infrastructure and services
NVIDIA GPUs
Infrastructure and services
Prometheus
Infrastructure and services
RabbitMQ
Infrastructure and services
Stripe
Infrastructure and services
System metrics
Infrastructure and services
Telegraf
Infrastructure and services
ZooKeeper
Languages and runtimes
.NET
Languages and runtimes
Browser
Languages and runtimes
C++
Languages and runtimes
Erlang/Elixir
Languages and runtimes
Go
Languages and runtimes
Java
Languages and runtimes
JavaScript / TypeScript
Languages and runtimes
Node.js
Languages and runtimes
PHP
Languages and runtimes
Python
Languages and runtimes
Ruby
Languages and runtimes
Rust
Languages and runtimes
Swift
Popular starting points
Bring the stack you already use.
Start with one service or agent. These are the integrations teams reach for first.
One connected system
Agents and everything they touch.
Logfire follows work across the agent runtime, application, and infrastructure, not just the model call.
Infrastructure
Web services, databases, queues, logs, cloud runtimes, and the calls between them.
FastAPI · PostgreSQL · Redis · AWS LambdaAgent frameworks
Model calls, tool use, messages, tokens, cost, and agent runs in one trace.
Pydantic AI · LangGraph · CrewAI · Vercel AI SDKLanguages
First-party SDKs for Python, TypeScript, and Rust; standard OTLP everywhere else.
Python · TypeScript · Rust · Go · Java · .NETPydantic AI Harness
Give long-running agents a workspace, plan, subagents, memory, and guardrails.
Filesystem · Planning · Compaction · GuardrailsNo rip and replace
Keep the backend you already have
OpenTelemetry can export the same signal to Logfire and your incumbent backend. Start with one agent or service, compare the traces, then move only what earns the move, without instrumenting the application twice.
Connect an existing OpenTelemetry stackStart with one service