#5151st of 89 in Agent platforms
kiln
Desktop app and Python library for AI task evals, optimization and fine-tuning
Live demo ↗
(opens in a new tab)Documentation ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 5.3k
- License
- custom license
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- Python
4,710 on 21 Apr 20265,271 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Kiln is a desktop app for Mac, Windows and Linux plus an MIT-licensed Python library (kiln-ai). One task and dataset flow through evals, prompt optimization, RAG, synthetic data, subagents and fine-tuning on Fireworks, Together and Vertex. It runs locally with your own API keys or offline with Ollama, and syncs team projects through Git.
Who it is for: Teams building and tuning LLM tasks who want evals and fine-tuning in one tool
Strengths
- Evals, RAG, synthetic data and fine-tuning share one dataset in a single tool
- Non-engineers rate outputs in the app; engineers load the same projects via pip
- Local-first: bring your own keys or run offline with Ollama; datasets are open JSON
- Team sync uses your existing Git infrastructure
Weaknesses
- Desktop app is source-available (fair-code); only the library and REST server are MIT
- AI Assistant, Auto-Optimize and Eval Builder require the paid Kiln Pro service
- Fine-tuning runs on hosted providers (Fireworks, Together, Vertex), not local GPUs
- No Docker or compose deployment; it is an installed desktop app
What it needs
- GPU optional
- Needs Ollama (optional, for offline use), Git (for team sync)
- Models: OpenAI, Anthropic, Gemini, Bedrock, Ollama
What changed
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in Agent platforms
See all 89| Rank | Project | Score |
|---|---|---|
| 1 | AutoGPTRole-based AI agents coordinated by one lead agent, plus a visual workflow builder | 95 out of 100 |
| 2 | daily_stock_analysisLLM-driven daily analysis of watchlist stocks across six markets, pushed to chat apps | 95 out of 100 |
| 3 | LangflowVisual builder for AI agents and workflows, exposed as API or MCP server | 94 out of 100 |
| #4 | deer-flowAgent harness running sub-agents, memory and sandboxes with pluggable skills | 93 out of 100 |
| #5 | DifyVisual builder for LLM workflows, RAG pipelines and agents, with APIs | 92 out of 100 |