Back to directory

Arkor

An alpha-stage TypeScript framework for fine-tuning open-weight LLMs with type-safe configs, a local Studio web UI, and managed GPUs.

Arkor

Arkor lets a TypeScript or Node developer fine-tune open-weight language models the way they ship the rest of their application: type-safe configuration files instead of a separate config format, and a local Studio web UI for starting and watching training runs. A scaffolding command provisions a throwaway anonymous workspace with no signup required, so a real training run can start immediately, with an account claimed later if desired.

Starter templates pair a small open-weight base model with a curated public dataset so the first run finishes in minutes; example templates cover support-ticket triage, translation, and PII redaction, each producing typed structured output. Training, evaluation and serving are intended to live in the same codebase and review flow as the rest of a product.

It is an alpha project whose APIs may change without notice. It is written in TypeScript and licensed under MIT.

More in Models

See all 405
  • mini-AGI

    github.com

    A continual learning language model that trains from scratch on a single 8 GB VRAM GPU, assembling its own architecture and adapting its size to the hardware and data available rather than shipping as a fixed pretrained model.

  • An open-source contrastive-embedding pipeline that maps technical CVE vulnerability descriptions to formal cybersecurity compliance controls in NIST SP 800-53 and CMMC, published with its model generation, fine-tuning and evaluation code.

  • peekaboolean

    github.com

    A small vision-language model, fine-tuned with LoRA on a 500-million-parameter backbone, that answers typed multiple-choice, ordinal-score and yes/no questions about an image with calibrated probabilities, in about 400 milliseconds on a laptop.