Skip to main content

Langfuse

Langfuse is an open-source LLM observability platform. It provides tracing and monitoring for AI workflows, helping teams debug, analyze, and optimize flow executions. Orbis Studio integrates with Langfuse out of the box via environment variables.

This guide covers two deployment scenarios:

  • Cloud — Langfuse Cloud, for customers running Orbis Studio on cloud infrastructure.
  • On-premises — Self-hosted Langfuse running as a service within the same Docker Compose network as Orbis Studio.

Prerequisites​

  • A running Orbis Studio server with at least one flow to trace.
  • Access to a Langfuse instance (Cloud or on-prem).

Cloud setup​

Use this setup when your Langfuse instance is hosted on Langfuse Cloud.

1. Obtain your Langfuse API keys​

In your Langfuse Cloud project, go to Settings → API Keys and create a new key pair. Copy the following:

  • Secret key (sk-lf-...)
  • Public key (pk-lf-...)
  • Base URL (for example, https://cloud.langfuse.com or https://us.cloud.langfuse.com)

2. Add credentials to your .env file​


_10
LANGFUSE_SECRET_KEY=sk-lf-...
_10
LANGFUSE_PUBLIC_KEY=pk-lf-...
_10
LANGFUSE_BASE_URL=https://cloud.langfuse.com

3. Restart Orbis Studio​


_10
uv run langflow run --env-file .env

Run a flow — Orbis Studio will automatically send trace data to your Langfuse Cloud project.

On-premises setup​

Use this setup when Langfuse is deployed as a service within the same Docker Compose network as Orbis Studio. The Langfuse service is accessible at http://langfuse:3000 within the compose network.

1. Obtain your Langfuse API keys​

Log in to your self-hosted Langfuse instance and go to Settings → API Keys to create a new key pair.

2. Add credentials to your Orbis Studio environment​

In the environment section of your Orbis Studio service in docker-compose.yml:


_10
environment:
_10
- LANGFUSE_SECRET_KEY=sk-lf-...
_10
- LANGFUSE_PUBLIC_KEY=pk-lf-...
_10
- LANGFUSE_BASE_URL=http://langfuse:3002

Because both services share the same Docker Compose network, Orbis Studio can reach Langfuse directly using the service name langfuse as the hostname. Langfuse listens on port 3002 in the Orbis Studio reference deployment.

3. Verify the connection​

After starting the stack, confirm that Orbis Studio can reach Langfuse:


_10
docker compose exec langflow python -c "import requests, os; addr = os.environ.get('LANGFUSE_BASE_URL'); print(addr); res = requests.get(addr, timeout=5); print(res.status_code)"

A 200 response confirms the connection is working.

Disable Langfuse tracing​

To disable the integration, remove the three LANGFUSE_* environment variables and restart Orbis Studio.