Integrations overview

Auto-instrument 60+ LLM providers, AI frameworks, and vector databases with a single openlit.init() call

OpenLIT uses OpenTelemetry auto-instrumentation to monitor your AI stack. When you call openlit.init(), the SDK automatically detects which supported libraries are installed in your environment and instruments them — no manual configuration required per library.

import openlit

# Instruments every supported library found in your environment
openlit.init()

import Openlit from 'openlit';

// Instruments every supported library found in your environment
Openlit.init();

Auto-instrumentation is non-invasive. OpenLIT patches library internals at startup using the OpenTelemetry instrumentation API. Your application code does not need to change.

Supported integrations

LLM providers

OpenAI, Anthropic, Cohere, Mistral, Groq, AWS Bedrock, Vertex AI, and 20+ more providers

AI frameworks

LangChain, LlamaIndex, CrewAI, OpenAI Agents, Pydantic AI, AG2, Haystack, and more

Vector databases

ChromaDB, Pinecone, Qdrant, Milvus, Astra DB, and PostgreSQL

LLM providers

Provider	Python	TypeScript
OpenAI	✅	✅
Anthropic	✅	✅
Cohere	✅	—
Mistral	✅	—
Groq	✅	—
AWS Bedrock	✅	—
Vertex AI	✅	—
Google AI Studio	✅	—
Azure OpenAI	✅	✅
Ollama	✅	✅
vLLM	✅	—
HuggingFace Transformers	✅	—
DeepSeek	✅	—
AI21	✅	—
Reka	✅	—
LiteLLM	✅	—
GPT4All	✅	—
Together AI	✅	✅
NVIDIA NIM	✅	—
xAI	✅	—
Sarvam	✅	—
PremAI	✅	—
Featherless	✅	—
Krutrim	✅	—
TitanML	✅	—
Julep	✅	—
MultiOn	✅	—
ElevenLabs	✅	—
AssemblyAI	✅	—

AI frameworks

Framework	Python	TypeScript
LangChain	✅	✅
LlamaIndex	✅	✅
CrewAI	✅	—
OpenAI Agents SDK	✅	—
Pydantic AI	✅	—
AG2 / AutoGen	✅	—
Haystack	✅	—
Mem0	✅	—
Browser Use	✅	—
Letta	✅	—
Crawl4AI	✅	—
Firecrawl	✅	—
Dynamiq	✅	—
ControlFlow	✅	—
SwarmZero	✅	—
DSPy	✅	—
Agno (Phidata)	✅	—
GuardrailsAI	✅	—

Vector databases

Database	Python	TypeScript
ChromaDB	✅	✅
Pinecone	✅	✅
Qdrant	✅	✅
Milvus	✅	—
Astra DB	✅	—
PostgreSQL (psycopg3)	✅	—

Disabling specific instrumentors

To skip instrumentation for one or more libraries, pass their names to the disabled_instrumentors parameter. Use the instrumentor key names shown in the table below.

import openlit

openlit.init(
    disabled_instrumentors=["cohere", "pinecone"]
)

import Openlit from 'openlit';

Openlit.init({
    disabledInstrumentations: ["cohere", "pinecone"]
});

You can also set this via the environment variable:

export OPENLIT_DISABLED_INSTRUMENTORS="cohere,pinecone"

Use disabled_instrumentors when you need fine-grained control — for example, to avoid double-instrumentation if you already use an upstream OpenTelemetry instrumentor for the same library.

---

GPU integrations

Monitor NVIDIA and AMD GPU utilization, memory, temperature, and power

Configuration

Full reference for all openlit.init() parameters

Destinations

Send telemetry to Datadog, Grafana, New Relic, and other observability stacks