# Arkor

> The TypeScript framework for fine-tuning open-weight LLMs.

## Docs

- [CLI](https://docs.arkor.ai/cli/overview.md): The Arkor CLI: dev, build, start, and friends.
- [Lifecycle callbacks](https://docs.arkor.ai/concepts/lifecycle.md): The five callbacks Arkor fires during a training run, and what each one is good for.
- [Concepts](https://docs.arkor.ai/concepts/overview.md): The mental model behind Arkor: project layout, trainer, lifecycle callbacks, and Studio.
- [Project structure](https://docs.arkor.ai/concepts/project-structure.md): Where Arkor expects your code, and what lives in .arkor/ and ~/.arkor/.
- [Studio](https://docs.arkor.ai/concepts/studio.md): The local web UI for starting training runs, watching them stream, and chatting with finished models.
- [Trainer](https://docs.arkor.ai/concepts/trainer.md): createTrainer: required fields, dataset sources, LoRA settings, and the start / wait / cancel lifecycle.
- [Introduction](https://docs.arkor.ai/introduction.md): Arkor is the TypeScript framework for fine-tuning open-weight LLMs.
- [Quickstart](https://docs.arkor.ai/quickstart.md): Scaffold an Arkor project, fine-tune a small open-weight LLM, and chat with it from a local Studio.
- [SDK](https://docs.arkor.ai/sdk/overview.md): The Arkor SDK: createArkor, createTrainer, infer, and the lifecycle callbacks.

## OpenAPI Specs

- [openapi](https://docs.arkor.ai/api-reference/openapi.json)