Send to OpenLIT

Configure the OpenLIT SDK to send traces and metrics to a self-hosted OpenLIT platform instance

OpenLIT ships with a built-in OpenTelemetry Collector. When you self-host the OpenLIT platform, it exposes an OTLP HTTP endpoint on port 4318 that accepts traces, metrics, and logs from your applications.

Get started

1. Deploy the OpenLIT platform

   If you haven’t deployed OpenLIT yet, follow the Installation Guide to set it up.

   Common platform endpoints:

   Deployment	Endpoint
   Local development	http://127.0.0.1:4318
   Docker Compose	http://localhost:4318
   Kubernetes	http://openlit.openlit.svc.cluster.local:4318
   External / Ingress	Your configured external endpoint

2. Install the OpenLIT SDK

   Python

   pip install openlit
   

   TypeScript

   npm install openlit
   
3. Configure the SDK to send to OpenLIT

   Python

   Using function arguments:

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

   Using environment variables:

   export OTEL_EXPORTER_OTLP_ENDPOINT="http://127.0.0.1:4318"
   export OTEL_SERVICE_NAME="my-ai-service"
   export OTEL_DEPLOYMENT_ENVIRONMENT="production"
   

   import openlit
   
   openlit.init()
   

   TypeScript

   Using function arguments:

   import Openlit from 'openlit';
   
   Openlit.init({
       otlpEndpoint: 'http://127.0.0.1:4318'
   });
   

   Using environment variables:

   export OTEL_EXPORTER_OTLP_ENDPOINT="http://127.0.0.1:4318"
   

   import Openlit from 'openlit';
   
   Openlit.init();
   

   Replace http://127.0.0.1:4318 with your actual OpenLIT platform endpoint if different.

No authentication required

By default, the OpenLIT platform’s OTLP endpoint does not require authentication headers. You only need to set otlp_endpoint.

If you expose the OpenLIT platform endpoint externally via an ingress or load balancer, configure authentication at the infrastructure level (e.g., mutual TLS or an API gateway) rather than in the SDK.

Full example

import openlit
import openai

openlit.init(
    otlp_endpoint="http://127.0.0.1:4318",
    service_name="my-ai-service",
    environment="production"
)

client = openai.OpenAI()
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)

Traces from this application appear in the OpenLIT platform under the service name my-ai-service.

---

Send to Grafana Cloud

Configure the Grafana OTLP gateway endpoint and Basic Auth headers

Send to Datadog

Route traces through the Datadog Agent to the Datadog backend

Integrations

60+ LLM providers, AI frameworks, and vector databases