Quick start

Installation Options

pipuv (isolated)pipx (isolated)From source/dev

Use pip to install mini in your current environment:

pip install mini-swe-agent

And try our command line interface

mini  # simple UI
mini -v  # visual UI
mini-extra  # extra utilities

Use uv/uvx (installation) to install & run the mini agent in an isolated environment.

Quickly install + run:

uvx mini-swe-agent  # simple UI
uvx mini-swe-agent -v  # visual UI
uvx --from mini-swe-agent mini-extra  # extra utilities

Permanently install

uv tool install mini-swe-agent
# then
mini  # simple UI
mini -v  # visual UI
mini-extra  # extra utilities

Use pipx (installation) to install & run mini in an isolated environment.

Quick install + run:

# Simple UI
pipx run mini-swe-agent
# Textual UI
pipx run mini-swe-agent -v
# Extra utilities
pipx run --spec mini-swe-agent mini-extra

or for a persistent installation (recommended):

pipx install mini-swe-agent
# then
mini  # simple UI
mini -v  # visual UI
mini-extra  # extra utilities

If the invocation doesn't immediately work, you might need to run pipx ensurepath.

For development or if you want to customize the agent:

git clone https://github.com/SWE-agent/mini-swe-agent.git
cd mini-swe-agent
pip install -e .

Then run:

mini  # simple UI
mini -v  # visual UI
mini-extra  # extra utilities

Or pick a run script:

python minisweagent/run/hello_world.py

If you are planning to contribute, please also install the dev dependencies and pre-commit hooks:

pip install -e '.[dev]'
pip install pre-commit && pre-commit install

To check your installation, you can run pytest -n auto in the root folder. This should run all tests in parallel (should take ~3min to run).

Note that there are still some extra dependencies that are not installed by default (basically anything that is in an .../extra/... folder). If you truly want to get the maximal package, you can run pip install -e '.[full]'

Changelog

Please see the github release notes for recent changes.

Example Prompts

Try mini-SWE-agent with these example prompts:

Implement a Sudoku solver in python in the sudoku folder. Make sure the codebase is modular and well tested with pytest.
Please run pytest on the current project, discover failing unittests and help me fix them. Always make sure to test the final solution.
Help me document & type my codebase by adding short docstrings and type hints.

Models

TLDR: Models should be set up the first time you run mini

If you missed the setup wizard, just run mini-extra config setup, or take a look at the following section.

Setting API keys

There are several ways to set your API keys:

Recommended: Run our setup script: mini-extra config setup. This should also run automatically the first time you run mini.
Use mini-extra config set ANTHROPIC_API_KEY <your-api-key> to put the key in the mini config file.
Export your key as an environment variable: export ANTHROPIC_API_KEY=<your-api-key> (this is not persistent if you restart your shell, unless you add it to your shell config, like ~/.bashrc or ~/.zshrc).
If you only use a single model, you can also set MSWEA_MODEL_API_KEY (as environment variable or in the config file). This takes precedence over all other keys.
If you run several agents in parallel, see our note about rotating anthropic keys here.

Selecting a model

All command line interfaces allow you to set the model name with -m or --model.
mini-extra config setup can set the default model for you
In addition, you can set the default model with mini-extra config set MSWEA_MODEL_NAME <model-name>, by editing the global config file (shortcut: mini-extra config edit), or by setting the MSWEA_MODEL_NAME environment variable.
You can also set your model in a config file (key model_name under model).

bug_report Something broken/unclear?

Open an issue on GitHub!

help Open-ended discussions

Join our Slack!

Our projects