AI Toolkit

license

MIT(OSI)

source

public(ostris/ai-toolkit)

core features withheld

no — no enterprise or ee path in the repository root

AI Toolkit is released under MIT, an OSI license, with the full source public and no enterprise, paid, or license-gated build. The README's Ostris Cloud section offers rented GPUs and an affiliate link to fund development, not a paid tier of the software.

• https://raw.githubusercontent.com/ostris/ai-toolkit/main/LICENSE recorded 2026-10-07

  LICENSE body, read in full: the stock MIT licence text, carrying no appended term or additional condition.

• https://raw.githubusercontent.com/ostris/ai-toolkit/main/README.md recorded 2026-10-07

  README, read for the distribution and gating question: it documents manual installation and running the code from this repository ("It can be run as a GUI or CLI"), and its Ostris Cloud section asks users to rent GPUs there to support the project ("please consider using [Ostris Cloud]"), beside a Runpod affiliate link. No enterprise, pricing or licence-gated build of the software appears.

No package registry tracks this product, so GitHub stars are the only comparable adoption signal available. A star is not a use, which limits how much the count can support and keeps confidence low.

• https://api.github.com/repos/ostris/ai-toolkit recorded 2026-08-31

  Repo metadata for ostris/ai-toolkit: stargazers_count 11860, archived false, pushed_at 2026-08-30.

AI Toolkit is a diffusion-specialist trainer with wide, current model coverage, but its methods are limited to LoRA and LyCORIS adapters and it has no distributed-training support. That puts it closer to single-GPU tools such as OneTrainer than to frontier-scale trainers like Megatron-LM.

• https://raw.githubusercontent.com/ostris/ai-toolkit/main/README.md recorded 2026-10-07

  README: "AI Toolkit is an easy to use all in one training suite for diffusion models". The "Supported Models" section lists FLUX.2-dev, FLUX.2-klein-base-4B and -9B, "Flex.1-alpha", "Flex.2-preview", "Chroma1-Base", "Lumina-Image-2.0" and Qwen-Image under Image, with Instruction / Edit, Video (Wan 2.1 and 2.2), Audio, LLM and Experimental groups; the README documents LoRA and LoKr training through LyCORIS and running as "a GUI or CLI".