AIエージェント
agent-device
Mobile app automation and verification built for AI coding agents — CLI, MCP server, and Node API
★ 4.9k スターMIT
ガイドを見る →deep-researchは、u14app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: u14app/deep-research）。GitHubスターは4,694、フォークは1,059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2026年6月18日です。
Deep Research with any LLM: in-depth reports in minutes, with an SSE API and MCP server built in
The recommended path is a free Gemini API key; alternatively configure any supported LLM and search provider.
Use the one-click Vercel/Cloudflare deploy, the Docker image, or a local dev install.
Set your LLM (and optional search) API keys in the environment file or settings.
Enter a topic, answer the clarifying questions, review the plan, and generate the report.
Topic → clarifying questions → research plan → SERP-driven collection → report, with the option to re-research from any stage.
Thinking/Task model split across many providers, and search via SearXNG, Tavily, Firecrawl, Exa, Bocha, Brave, or the model itself.
Upload text, Office, and PDF files to ground research in your own documents.
Edit reports in WYSIWYM or Markdown modes, adjust reading level and length, translate the full text, and generate a knowledge graph with one click.
Expose research as a service: a streaming SSE API and an MCP server (Streamable HTTP and SSE transports) for other AI tools.
Research data stays in local browser storage; deploy to Vercel, Cloudflare, Docker, or as a static export, and install it as a PWA.
Clone the project and enter its directory.
git clone https://github.com/u14app/deep-research.git
cd deep-researchInstall dependencies with pnpm and copy the environment template (Node.js 18.18+ recommended).
pnpm install
cp env.tpl .env.localStart the app and open it in your browser; add your LLM API key in .env.local or the UI first.
pnpm dev
# then open http://localhost:3000The u14app team; a hosted instance runs at research.u14.app.
The software is MIT-licensed and free. You bring your own model access — a free Gemini API key is the recommended path, and local models via Ollama avoid per-call costs.
It needs an LLM; Gemini's free tier works per the README, and search providers are optional since some models and SearXNG can cover search.
It offers a similar staged research workflow but is open source, model-agnostic, self-hostable, and keeps data in your browser — with API and MCP access the hosted products don't offer in the same way.
Get a Gemini API key, use the one-click Vercel deploy or run `pnpm install && pnpm dev` locally, then start a research topic at http://localhost:3000.
Mobile app automation and verification built for AI coding agents — CLI, MCP server, and Node API
The unofficial Home Assistant MCP server — control and automate your smart home from any AI assistant.
An autonomous research agent that plans, searches and writes cited reports from any data source
MCP server letting AI agents run 150+ cybersecurity tools for authorized pentesting and research.
Brave's official MCP server: web, news, image, video, local, and LLM-context search for your agents.
Parallel multi-agent research that compiles wikis from any topic.
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 — deep-research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deep-research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。
Rivet is Ironclad's visual IDE and TypeScript library for building complex AI agents and prompt chains
Koog is a JVM (Java and Kotlin) framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend service…
A team workspace for building AI agents together.
Microsoft's open framework for red-teaming generative AI systems.
Ghidra MCP server with 200+ tools that let AI assistants drive real reverse-engineering work
Convene 18 legendary thinkers as an AI council that debates your hardest questions — inside Claude Code