AIエージェント
patchwork
Agentic AI framework for automating enterprise workflows like code review.
★ 1.6k スターAGPL-3.0
ガイドを見る →devikaは、stitionaiが公開しているオープンソースのAIエージェントプロジェクトです（GitHub: stitionai/devika）。GitHubスターは19,555、フォークは2,566で、MITライセンスの下で公開されています。GitHubでの最終更新は2025年9月25日です。
Devika - the open-source agentic software engineer that plans, browses and writes code
Write the feature or small app you want Devika to build.
Get the repository from GitHub.
Create a uv virtual environment, install the Python requirements, and install Playwright browsers.
Launch the backend, open the web UI, configure your model keys, and submit your goal.
Breaks a goal into steps and reasons about them before writing code.
Works with Claude 3, GPT-4, Gemini, Mistral, Groq, or local models through Ollama.
Browses the web and extracts contextual keywords to research focused information for the task.
Generates project code across multiple programming languages.
Tracks and visualizes the agent's state dynamically as it works.
Organizes work into projects you manage through the chat-based web interface.
Get the source code from GitHub and enter the project directory.
git clone https://github.com/stitionai/devika.git
cd devikaDevika uses uv for its Python environment, plus Playwright for browsing.
uv venv
uv pip install -r requirements.txt
playwright install --with-depsStart the backend entry point, then open the web interface it serves.
python devika.pyIt is maintained by stitionai and contributors on GitHub.
Yes - Devika is MIT licensed and free to use.
With cloud models you need your own API keys (Anthropic, OpenAI, Google, Mistral or Groq). Running local models through Ollama avoids API costs.
Compared with Devin, Devika is the open-source, self-hosted take on the agentic software engineer - you bring your own models and infrastructure, with a smaller but inspectable feature set.
Clone the repo, install requirements with uv, run python devika.py, and give it a small project in the web UI.
Agentic AI framework for automating enterprise workflows like code review.
An autonomous research agent that plans, searches and writes cited reports from any data source
smolagents — Hugging Face's barebones agent library where agents think in code: a CodeAgent that writes its actions as Python, model-agnostic, Apache-2.0
Local-first agentic workspace where Claude Code, Codex and a built-in agent share memory, apps and tools.
An agentic skills framework & software development methodology that works.
AI full-stack engineer agent built on Next.js, Claude, and Kubernetes.
LibreAIカタログに掲載されたAIエージェントプロジェクトの中で最もスターの多いもの — devikaが合わなかった場合に検討したい、現実的な候補リストです。
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
カタログの中でdevikaとスター数が近いAIエージェントプロジェクト — 前後の棚もそのまま見て回れます。
NVIDIA's security scanner for AI agent skills - check before you install
Open source agentic operating system
Tongyi Deep Research - Alibaba's open-source deep research agent and 30B model
Agent Zero AI framework
Kortix - the open-source AI operating system: agents, skills and memory as files in a repo
Unit tests for LLMs — evaluate before you ship.