voice-ai

Unsloth: Train AI Models

Train and run open AI models locally with Unsloth Studio.
intermediate⏱ 30 minutesπŸ’΅ Free
75,388 stars6,844 forksPythonGuide quality 8/10Updated 9/1/2026100% free Β· open source
What it is

Builds a web UI for training and running open models like Gemma 4 locally.

Real use case example

"A founder uses Unsloth Studio to build a basic chat app that responds to users' queries, training a model overnight and deploying it in the morning."

Best insideClaude DesktopSelf-hostedAny AI Client
When to use it
  • β€’When you want to use open AI models without relying on cloud services
  • β€’When you need to train AI models with sensitive data that cannot be shared publicly
  • β€’When you want to reduce the latency and costs associated with cloud-based AI model deployment
Ready-to-paste prompt
Run `python train.py --model gemma4 --train_data_path /path/to/train/data` to fine-tune the Gemma 4 model with your custom dataset
Heads up: Make sure you have the required Python version (3.8 or higher) and the necessary dependencies installed, including `transformers` and `torch`, before running Unsloth
Saves to your device
Use with Claude
New

Skip the builder β€” one click puts this in Claude, Cursor, Antigravity and more.

πŸ› οΈ Technical setup

Expect 20–40 minutes in a terminal β€” or let your AI agent drive it.

Try it instantly β€” no install
Claude Code
mkdir -p ~/.claude/skills/unsloth && curl -fsSL https://workflowstacks.com/api/skills/unsloth/claude-skill -o ~/.claude/skills/unsloth/SKILL.md
Open in another AI app

Opens the app with this repo with the prompt ready to go β€” no copy-paste needed.

Connect the whole catalog (MCP)
claude mcp add --transport http workflowstacks https://workflowstacks.com/api/mcp

Adds a WorkflowStacks connector to Claude Code: search and load any skill here by chatting.

How Unsloth: Train AI Models works
Codeflow
Free to inspect

Unsloth: Train AI Models is a very large Python project (~684k lines across 1873 code files, plus 1559 test files). It is a full software project: use it through its install path rather than reading it end to end. Last commit this month, Apache-2.0 license, has a test suite.

Size
Very large codebase
~684k lines Β· 1873 code files Β· days to read β€” use, don't read
Setup
Developer setup
A real software project. Use it via its install path; don't expect to read it all.
Runs on
Python
No API keys detected
Python 72%TypeScript 22%Shell 2%PowerShell 2%Rust 2%
What happens, step by step

Train AI models locally

You give
AI model data
1Install Tool
Download and install Unsloth
install.sh
2Load Model
Load AI model for training
3Train Model
Train AI model locally
cli.py
4Deploy Model
Export or deploy trained model
unsloth_cli/__main__.py
You get
Trained model
Where to start reading
  1. 1
    README.md
    Start here β€” what it does and how to install it
  2. 2
    cli.py
    Where the program starts running
  3. 3
    pyproject.toml
    Dependencies and the commands it exposes
  4. 4
    unsloth/__init__.py
    Inside unsloth/ β€” the main logic begins here
What's in each folder
unsloth/Core code β€” the actual logic101 files
studio/frontend/Frontend / UI1707 files
studio/backend/Backend / API1183 files
studio/src-tauri/Folder55 files
scripts/Helper scripts53 files
unsloth_cli/Folder39 files
tests/Tests β€” proof it works551 files
.github/CI / automation (GitHub Actions)83 files
READMEHas testsDocumentedCI checksApache-2.0 licenseUpdated this month
Quick Actions
Details
Creator
unslothai
Language
Python
Category
voice-ai
Published
11/29/2023

Are you the creator of this tool? Claim your listing β†’ and earn 85% of every sale.