https://github.com/Kiln-AI/Kiln
The easiest tool for fine-tuning LLM models, synthetic data generation, and collaborating on datasets.
https://github.com/Kiln-AI/Kiln
ai chain-of-thought collaboration dataset-generation fine-tuning machine-learning macos ml ollama openai prompt prompt-engineering python rlhf synthetic-data windows
Last synced: about 1 month ago
JSON representation
The easiest tool for fine-tuning LLM models, synthetic data generation, and collaborating on datasets.
- Host: GitHub
- URL: https://github.com/Kiln-AI/Kiln
- Owner: Kiln-AI
- Created: 2024-07-23T23:10:13.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-01-22T06:53:52.000Z (10 months ago)
- Last Synced: 2025-01-23T14:07:45.247Z (10 months ago)
- Topics: ai, chain-of-thought, collaboration, dataset-generation, fine-tuning, machine-learning, macos, ml, ollama, openai, prompt, prompt-engineering, python, rlhf, synthetic-data, windows
- Language: Python
- Homepage: https://docs.getkiln.ai
- Size: 3.04 MB
- Stars: 808
- Watchers: 13
- Forks: 36
- Open Issues: 14
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
Awesome Lists containing this project
- awesome - Kiln-AI/Kiln - The easiest tool for fine-tuning LLM models, synthetic data generation, and collaborating on datasets. (Python)
- StarryDivineSky - Kiln-AI/Kiln - AI/Kiln 是一个易于使用的工具，用于微调大型语言模型（LLM），生成合成数据，以及协作处理数据集。它旨在简化LLM模型的定制和数据管理流程。Kiln可能提供友好的用户界面或API，方便用户上传、标注和处理数据。通过微调，用户可以使LLM模型更适应特定任务或领域。合成数据生成功能可以帮助用户扩充数据集，解决数据稀缺问题。协作功能则方便团队成员共同参与数据处理和模型训练过程。Kiln的目标是降低LLM技术的使用门槛，让更多人能够利用LLM解决实际问题。具体工作原理和技术细节需要进一步研究项目代码和文档。 (A01_文本生成_文本对话 / 大语言对话模型及数据)
- awesome-repositories - Kiln-AI/Kiln - The easiest tool for fine-tuning LLM models, synthetic data generation, and collaborating on datasets. (Python)
- awesome-LLM-resources - Kiln - tuning LLM models, synthetic data generation, and collaborating on datasets. (微调 Fine-Tuning)
- definitive-opensource - Kiln - tuning LLM models, synthetic data generation, and collaborating on datasets. | `Cross` | **4.4k** | (Table of Contents / Model Tools)
- awesome-llm-agents - Kiln AI - Tools for building AI products including (Frameworks)
- awesome-mcp-servers - **Kiln** - A free open-source platform for building production-ready AI systems. It supports RAG pipelines, AI agents, MCP tool-calling, evaluations, synthetic data generation, and fine-tuning — all in one unified framework by Kiln-AI. `http` `ai` `git` `github` (📦 Other)
README
Easily Build AI Systems
Kiln is a free and intuitive app for building AI systems and products
Evals •
RAG •
Agents •
Fine Tuning •
Synthetic Data •
All Docs
## Key Features
- 🚀 [**Intuitive Desktop Apps**](https://kiln.tech/download): One-click apps for Windows, MacOS, and Linux.
- 📊 [**Evals**](https://docs.kiln.tech/docs/evaluations): Evaluate the quality of your models/tasks using state of the art evaluators.
- 🎛️ [**Fine Tuning**](https://docs.kiln.tech/docs/fine-tuning-guide): Zero-code fine-tuning for Llama, GPT-4o, and more. Automatic serverless deployment of models.
- 🔍 [**Docs & Search (RAG)**](https://docs.kiln.tech/docs/documents-and-search-rag): Add knowledge to your AI systems with Retrieval-Augmented Generation (RAG).
- 🤖 [**Agents**](https://docs.kiln.tech/docs/agents): Build agentic systems with multiple actors
- 🛠 [**Tools & MCP**](https://docs.kiln.tech/docs/tools-and-mcp): Connect powerful tools to your Kiln tasks
- 🪄 [**Synthetic Data Generation**](https://docs.kiln.tech/docs/synthetic-data-generation): Generate eval datasets or fine-tuning data with our interactive visual tooling.
- 🧠 [**Reasoning Models**](https://docs.kiln.tech/docs/guide-train-a-reasoning-model): Train or distill your own custom reasoning models.
- 📝 [**Prompt Generation**](https://docs.kiln.tech/docs/prompts): Automatically generate prompts including chain-of-thought, few-shot, and multi-shot, and more.
- 🌐 [**Comprehensive Model Support**](https://kiln.tech/model_library): Skip the guesswork — we've tested over 100 models' capabilities. Use any model via Ollama, OpenAI, OpenRouter, Fireworks, Groq, AWS, any OpenAI compatible API, and more.
- 🤝 [**Team Collaboration**](https://docs.kiln.tech/docs/collaboration): Git-based version control for your AI datasets. Intuitive UI makes it easy to collaborate with QA, PM, and subject matter experts on data samples, evals, prompts, ratings and issues.
- 🗃️ [**Structured Data**](https://docs.kiln.tech/docs/structured-data-json): Build AI tasks that speak JSON.
- 🧑💻 **Open-Source [Library](https://docs.kiln.tech/developers/python-library-quickstart) and [API](https://docs.kiln.tech/developers/rest-api)**: Our Python library and OpenAPI REST API are MIT open source.
- 🔒 [**Privacy-First**](https://docs.kiln.tech/docs/privacy): Kiln runs locally on your computer. We can't access your data. Bring your own API keys or use Ollama.
- 📚 [**Awesome Docs**](https://docs.kiln.tech): easy-to-follow video guides, approachable for beginners, and depth for advanced users.
- 💰 **Free**: Our apps are free and our library is open-source.
## Demo
[Watch a 2 minute overview of Kiln](https://kiln.tech#demo) or our [end to end project demo (20 minutes)](https://docs.kiln.tech/docs/end-to-end-project-demo).
## People in Our Community Work For
For privacy, Kiln doesn't track the identity of who uses it. People from these companies have joined our communities on Github & Discord.
## Download Kiln Desktop Apps
Available on MacOS, Windows and Linux.
[](https://kiln.tech/download)
## Docs & Guides
Kiln is quite intuitive, so we suggest launching the desktop app and diving in. However if you have any questions or want to learn more, our [docs are here to help](https://docs.kiln.tech).
### Video Guides
- [Fine Tuning LLM Models](https://docs.kiln.tech/docs/fine-tuning-guide)
- [Guide: Train a Reasoning Model](https://docs.kiln.tech/docs/guide-train-a-reasoning-model)
- [LLM Evaluators](https://docs.kiln.tech/docs/evaluators)
- [End to End Project Demo](https://docs.kiln.tech/docs/end-to-end-project-demo)
- [Tools 101: Intro to Tools & MCP](https://docs.kiln.tech/docs/tools-and-mcp)
### All Docs
- [Quick Start](https://docs.kiln.tech/getting-started/quickstart)
- [How to use any AI model or provider in Kiln](https://docs.kiln.tech/docs/models-and-ai-providers)
- [Documents & Search Tools (RAG)](https://docs.kiln.tech/docs/documents-and-search-rag)
- [Tools & MCP](https://docs.kiln.tech/docs/tools-and-mcp)
- [Reasoning & Chain of Thought](https://docs.kiln.tech/docs/reasoning-and-chain-of-thought)
- [Synthetic Data Generation](https://docs.kiln.tech/docs/synthetic-data-generation)
- [Collaborating with Kiln](https://docs.kiln.tech/docs/collaboration)
- [Rating and Labeling Data](https://docs.kiln.tech/docs/reviewing-and-rating)
- [Prompt Styles](https://docs.kiln.tech/docs/prompts)
- [Structured Data / JSON](https://docs.kiln.tech/docs/structured-data-json)
- [Organizing Kiln Datasets (Tags and Filters)](https://docs.kiln.tech/docs/organizing-datasets)
- [Our Data Model](https://docs.kiln.tech/docs/kiln-datamodel)
- [Repairing Responses](https://docs.kiln.tech/docs/repairing-responses)
- [Keyboard Shortcuts](https://docs.kiln.tech/docs/keyboard-shortcuts)
- [Privacy Overview: Private by Design](https://docs.kiln.tech/docs/privacy)
For developers, see our [Kiln Python Library Docs](https://kiln-ai.github.io/Kiln/kiln_core_docs/kiln_ai.html). These include how to load datasets into Kiln, or using Kiln datasets in your own code-base/notebooks.
## Build & Tools
| | |
| ------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| CI | [](https://github.com/Kiln-AI/kiln/actions/workflows/build_and_test.yml) [](https://github.com/Kiln-AI/kiln/actions/workflows/format_and_lint.yml) [](https://github.com/Kiln-AI/kiln/actions/workflows/build_desktop.yml) [](https://github.com/Kiln-AI/kiln/actions/workflows/web_format_lint_build.yml) |
| Tests | [](https://github.com/Kiln-AI/kiln/actions/workflows/test_count.yml) [](https://github.com/Kiln-AI/kiln/actions/workflows/test_count.yml) |
| Package | [](https://pypi.org/project/kiln-ai/) [](https://pypi.org/project/kiln-ai/) |
| Meta | [](https://github.com/astral-sh/uv) [](https://github.com/astral-sh/ruff) [](https://github.com/microsoft/pyright) [](https://kiln-ai.github.io/Kiln/kiln_core_docs/index.html) |
| Apps | [](https://kiln.tech/download) [](https://kiln.tech/download) [](https://kiln.tech/download)  |
| Connect | [](https://kiln.tech/discord) [](https://kiln.tech/blog) |
## Python Library
[](https://pypi.org/project/kiln-ai/) [](https://kiln-ai.github.io/Kiln/kiln_core_docs/index.html)
Our open-source [python library](https://pypi.org/project/kiln-ai/) allows you to integrate Kiln datasets into your own workflows, build fine tunes, use Kiln in Notebooks, build custom tools, and much more! [Read the docs](https://kiln-ai.github.io/Kiln/kiln_core_docs/index.html) for examples.
```bash
pip install kiln-ai
```
## Learn More
### Rapid Prototyping
There are new models and techniques emerging all the time. Kiln makes it easy to try a variety of approaches and compare them in a few clicks without writing code. These can result in higher quality and improved performance.
We currently support:
- Various prompting techniques: basic, few-shot, multi-shot, repair & feedback
- Chain of thought / thinking, with optional custom “thinking” instructions
- Many models: GPT, Llama, Claude, Gemini, Mistral, Gemma, Phi
- Fine Tuning: create custom models using your Kiln dataset
- Evaluations using LLM-as-Judge and G-Eval
- Distilling models
In the future, we plan to add more powerful no-code options like RAG. For experienced data-scientists, you can create these techniques today using Kiln datasets and our Python library.
### Collaborate Across Technical and Non-Technical Teams
When building AI products, there’s usually a subject matter expert who knows the problem you are trying to solve, and a different technical team assigned to build the model. Kiln bridges that gap as a collaboration tool.
Subject matter experts can use our intuitive desktop apps to generate structured datasets and ratings, without coding or using technical tools. No command line or GPU required.
Data-scientists can consume the dataset created by subject matter experts, using the UI, or deep dive with our python library.
QA and PM can easily identify issues sooner and help generate the dataset content needed to fix the issue at the model layer.
The dataset file format is designed to be used with Git for powerful collaboration and attribution. Many people can contribute in parallel; collisions are avoided using UUIDs, and attribution is captured inside the dataset files. You can even share a dataset on a shared drive, letting completely non-technical team members contribute data and evals without knowing Git.
### Build High Quality AI Products with Datasets
Products don’t naturally have “datasets”, but Kiln helps you create one. Every time you use Kiln, we capture the inputs, outputs, human ratings, feedback, and repairs needed to build high quality models for use in your product. The more you use it, the more data you have.
Our synthetic data generation tool can build datasets for evals and fine-tuning in minutes.
Your model quality improves automatically as the dataset grows, by giving the models more examples of quality content (and mistakes). If your product goals shift or new bugs are found (as is almost always the case), you can easily iterate the dataset to address issues.
## Contributing & Development
See [CONTRIBUTING.md](CONTRIBUTING.md) for information on how to setup a development environment and contribute to Kiln.
## Citation
```bibtex
@software{kiln_ai,
title = {Kiln: Rapid AI Prototyping and Dataset Collaboration Tool},
author = {{Chesterfield Laboratories Inc.}},
year = {2025},
url = {https://github.com/Kiln-AI/Kiln},
version = {latest}
}
```
## Licenses & Trademarks
- Python Library: [MIT License](libs/core/LICENSE.txt)
- Python REST Server/API: [MIT License](libs/server/LICENSE.txt)
- Desktop App: free to download and use under our [EULA](app/EULA.md), and [source-available](/app). [License](app/LICENSE.txt)
- The Kiln names and logos are trademarks of Chesterfield Laboratories Inc.
Copyright 2024 - Chesterfield Laboratories Inc.