Show HN: Karotte, a framework for building robust RL envs

Preference Model released Karotte, an open-source framework for building robust reinforcement learning environments, installable via `uv tool install karotte` and requiring Python 3.12+. Karotte runs tasks in a VM by default — Apple `container` on macOS and Firecracker on Linux, with docker or podman also supported — and writes each run's transcript to `out/transcript.json` for viewing with `karotte dashboard out/`. The framework is MIT licensed, while its environment templates in `src/karotte/templates/` use MIT No Attribution (MIT-0), and every merge to `main` is released.

Karotte is an open-source framework for building robust RL environments, made by Preference Model https://preferencemodel.com . Documentation: karotte.dev https://karotte.dev Karotte needs Python 3.12+ and uv https://docs.astral.sh/uv/ . Runs go into a VM by default: Apple container on macOS, Firecracker on Linux. docker or podman work too. See Installation https://karotte.dev/getting-started/installation/ and Runtimes https://karotte.dev/running/runtimes/ for what each needs. uv tool install karotte Create an environment from the default template and run its example task: karotte create-env my env cd my env uv sync --extra dev uv run setup data.py export ANTHROPIC API KEY=... uv run karotte run --task example-task --model anthropic/claude-opus-5-5 karotte run builds the image, runs the task, and writes the transcript to out/transcript.json . karotte dashboard out/ shows it. The quick start https://karotte.dev/getting-started/quick-start/ explains each step, and Writing tasks https://karotte.dev/tasks/tasks-and-steps/ shows how to add your own. just lint just test just test-template default Every merge to main is released. See Developing Karotte https://karotte.dev/extending/developing-karotte/ . Karotte is under the MIT license https://github.com/preferencemodel/karotte/blob/main/LICENSE . The templates in src/karotte/templates/ are under MIT No Attribution https://github.com/preferencemodel/karotte/blob/main/src/karotte/templates/LICENSE MIT-0 , so environments created from them need no license notice. Images built from the templates contain third-party software under its own licenses. They are based on Amazon Linux 2023. The language-toolchains template installs GPL and LGPL software gcc, GnuCOBOL, GNU Prolog, Free Pascal, the libraries bundled with Julia, and others , Amazon Corretto GPL-2.0 with the Classpath Exception and Clojure EPL-1.0 for the languages you enable. Mojo, off by default, has its own license terms; check them for the version you enable. If you distribute a built image, you're responsible for complying with the licenses of the software in it.