Are you an LLM? Read llms.txt for a summary of the docs, or llms-full.txt for the full context.
Skip to content

Mubit

Mubit passes your app's calls on to your model provider and records them. It groups similar calls into tasks and trains adapters, small models each made for one task. Each adapter version is tested against your provider model's answers, and you can compare a passing version with your provider model in Try models.

Give your client Mubit's base URL and your Mubit key:

import os
from openai import OpenAI
 
client = OpenAI(
    base_url="https://api.mubit.ai/openai/v1",
    api_key=os.environ["MUBIT_API_KEY"],
)

Your model IDs stay the same.

How it works

  1. Connect: give your client Mubit's base URL and your Mubit key.
  2. See your calls: each call shows in Requests, with its request and reply.
  3. Find tasks: calls with the same model, instructions, tools, tool choice and output format form a task.
  4. Train and test an adapter: Mubit tests each adapter version against your provider model's answers, and you can compare a passing version in Try models.

Concepts explains each term.

Start from

Good to know

  • Calls run on your provider key, or on the starter credit where it's offered. See Provider keys.
  • The examples use the host https://api.mubit.ai. Use the base URL your console's Connect page shows.