Quickstart

Get LLM observability running in under 5 minutes

This guide walks you through deploying OpenLIT and sending your first LLM trace.

1. Deploy OpenLIT

   Clone the repository and start the full stack with Docker Compose:

   git clone [e-mail]:openlit/openlit.git
   cd openlit
   docker compose up -d
   

   This starts:

   • OpenLIT UI at http://127.0.0.1:3000
   • OpenTelemetry Collector on ports 4317 (gRPC) and 4318 (HTTP)
   • ClickHouse database on port 8123
2. Install the OpenLIT SDK

   Python

   pip install openlit
   
3. Initialize OpenLIT in your application

   Add these two lines at the start of your application, before any LLM calls:

   Python

   import openlit
   
   openlit.init(otlp_endpoint="http://127.0.0.1:4318")
   

   If you omit otlp_endpoint, OpenLIT prints traces to the console — useful during development without a running collector.

4. Make an LLM call

   OpenLIT auto-instruments any supported LLM library you have installed. Here’s an example using OpenAI:

   Python

   import openlit
   from openai import OpenAI
   
   openlit.init(otlp_endpoint="http://127.0.0.1:4318")
   
   client = OpenAI()
   
   response = client.chat.completions.create(
       model="gpt-4o",
       messages=[{"role": "user", "content": "What is OpenTelemetry?"}]
   )
   print(response.choices[0].message.content)
   
5. View your traces

   Open http://127.0.0.1:3000 and log in with the default credentials:

   Field	Value
   Email	[e-mail]
   Password	openlituser

   Navigate to Requests to see your LLM traces, including model name, token usage, cost, and latency.

What gets traced automatically

OpenLIT auto-detects and instruments any supported library installed in your environment:

• LLM calls — model, prompt, completion, tokens, cost, latency, finish reason
• Agent traces — tool calls, chain steps, memory operations
• Vector DB operations — collection, operation type, latency
• GPU metrics — utilization, memory, temperature (when collect_gpu_stats=True)

No code changes are needed beyond openlit.init().

Next steps

SDK configuration

Full list of openlit.init() parameters for customizing instrumentation

Integrations

Browse all 50+ supported LLM providers, frameworks, and vector databases

Evaluations

Set up automatic LLM response quality evaluation

Dashboards

Build custom dashboards for your AI telemetry