LLM observability providers

An observability provider traces the LLM calls made by the Gen AI orchestrator (prompts, answers, latency, cost). It is configured in the Observability settings menu.

Here are the LLM observability providers supported by Tock:

Provider	provider value
Langfuse	Langfuse

Langfuse

{
  "provider": "Langfuse",
  "url": "http://localhost:3000",
  "public_url": "https://langfuse.example.com",
  "secret_key": {
    "type": "Raw",
    "secret": "sk-lf-****************-ceabe45abe8f"
  },
  "public_key": "pk-lf-****************-b77e68ef7d2c"
}

Field	Description
url	URL used by the orchestrator to reach the Langfuse server (default: http://localhost:3000)
public_url	Optional URL of the Langfuse server as seen from the users' browser, used for the links to the traces in Tock Studio. Defaults to url
secret_key	Langfuse secret key
public_key	Langfuse public key

Chat with Tock

×