Getting StartedInstallation

Installation

Install from Source

Clone the repository and install in development mode:

git clone https://github.com/MindLab-Research/mindlab-toolkit.git
cd mindlab-toolkit
pip install -e .

What’s Included

The MinT Python SDK provides low-level operations including forward_backward, sample, optim_step, and save_state functionality.

Authentication

API access requires setting up credentials:

  1. Obtain an API key from the Mind Lab team
  2. Set the MINT_API_KEY environment variable
export MINT_API_KEY=sk-mint-your-key-here

Tinker Compatibility

MinT targets API compatibility with Tinker (baseline: tinker 0.6.3). If you already have the tinker package installed, you can use it with MinT by setting environment variables:

export TINKER_BASE_URL=https://mint.macaron.im/
export TINKER_API_KEY=<your-mint-api-key>

Most code using import tinker will work with MinT without modification. For known gaps and planned updates, see Tinker Compatibility.

SDK Version Requirement

MinT API keys (sk-mint-*) require tinker SDK version 0.6.3:

pip install tinker==0.6.3

Newer SDK versions (0.8.0+) use a different key format (tml-*) and are not compatible with MinT API keys. If you see the error The api_key must start with the 'tml-' prefix, downgrade to tinker 0.6.3.

Note: The zero_grad_async() method does not exist in tinker 0.6.3. This is expected - gradient zeroing is handled automatically by the server.

Quickstart

For a hands-on tutorial, see the mint-quickstart repository which demonstrates SFT and RL training workflows.

IntroductionNavigating These Docs