Unsloth (open-source library)
Project record
Unsloth is an open-source Python library for fine-tuning and reinforcement learning of language, diffusion, text-to-speech, and embedding models, supporting LoRA, QLoRA, full fine-tuning, pretraining, GRPO, and DPO, and exporting to formats such as GGUF. The same repository contains Unsloth Studio, a web interface, and the Unsloth Desktop app.1
- Repository: Repository (external site: github.com)
- Documentation: Documentation (installation) (external site: unsloth.ai)
- License: LICENSE (Apache 2.0 core) (external site: github.com)
- License: COPYING (AGPL-3.0 components) (external site: github.com)
- Release notes: Releases (external site: github.com)
Availability and license
Overall availability
Documented as available to the general public. Access conditions and license terms may still apply.12
Availability is separate from permission: read the license before using or redistributing.
Apache License 2.0 (unsloth/, tests/, scripts/) (external site: github.com)2
GNU Affero General Public License v3.0 only (studio/, unsloth_cli/) (external site: github.com)234
The LICENSE file applies Apache 2.0 to the core package directories and the AGPL v3 to the optional Studio and command-line directories; source files in those directories carry the SPDX identifier AGPL-3.0-only. The PyPI metadata for the unsloth package declares Apache-2.0.246
Component reuse rights
- weights
- Unknown — no complete fact-level rights review
- code
- Unknown — no complete fact-level rights review
- data
- Unknown — no complete fact-level rights review
- documentation
- Unknown — no complete fact-level rights review
No complete system-rights review is recorded for this release.
Public materials checklist
| Item | Status | Notes and evidence |
|---|---|---|
| Source codeIs the source code publicly readable? | Public | 1 |
| DocumentationIs user documentation published? | Public | Installation and usage guides on unsloth.ai/docs.7 |
| InstallationAre installation instructions or packages publicly available? | Public | Documented installs by install script (macOS, Linux, WSL, and Windows PowerShell), pip or conda for Unsloth Core, Docker images (including a separate ROCm image), and desktop app downloads.17 |
| Supported platformsAre supported operating systems or hardware documented? | Public | The README lists Windows, Linux, WSL, and macOS, with NVIDIA, AMD, and Intel GPUs, CPUs, multi-GPU setups, and a Vulkan backend.1 |
| Release statusAre versioned releases published? | Public | Releases are published frequently on GitHub; the unsloth package version on PyPI at review was 2026.9.12, uploaded on September 28, 2026.56 |
What it is useful for
Fine-tuning and post-training open-weight models on a single machine or multi-GPU setup, and exporting the results for local inference.1
Organization context
U.S. eligibility
Eligible · basis: U.S.-governed project
The maintaining entity is Unsloth AI Inc.: the library is developed in the unslothai GitHub organization, and the repository's license and source-file copyright notices name the Unsloth AI Inc. team. The company's U.S. principal place of business is documented in an official registry: the USPTO registration of the UNSLOTH mark (No. 7973599), for downloadable software for training and fine-tuning large language models, names Unsloth AI Inc., a Delaware corporation, as owner with a San Francisco address; the Unsloth organization record explains why the trademark file supports reading this as the company's principal place of business. No foundation or multi-organization governance is documented.2489
Sources
This listing is not an endorsement, a safety assessment, or a federal approval.