The Data Layer for AI Systems
Transform 18 source types — docs, GitHub repos, PDFs, videos, notebooks, wikis, and more — into AI skills and RAG knowledge. One tool for Claude, Gemini, OpenAI, LangChain, Cursor, and 12+ AI platforms.
$ pip install skill-seekers
18 Sources → Any AI System
One tool for AI skills and RAG from any source
Input Sources
Documentation
Any doc site, with SPA browser rendering
GitHub Repos
Public & private repos, with AST analysis
PDF Files
Scanned docs, manuals, research papers with OCR
Local Codebases
27+ languages, C3.x deep analysis
Video
YouTube, local videos, with OCR code extraction
Jupyter Notebooks
Code cells, outputs, markdown extraction
All 18 source types supported via unified create command and MCP
12+ AI Platform Outputs
AI Skills
RAG / Vectors
AI Coding
Generic
The Data Layer for AI Systems
Skill Seekers transforms 18 source types — documentation sites, GitHub repos, PDFs, videos, Jupyter notebooks, Word/EPUB documents, OpenAPI specs, Confluence wikis, Notion pages, and more — into structured AI skills and RAG-ready knowledge for 12+ AI platforms.
The Problem
- × Building AI skills takes days of preprocessing — scraping, analyzing code, extracting patterns
- × AI assistants lack deep expertise without manual context preparation
- × Understanding new codebases requires weeks of manual analysis
- × Different AI systems need different formats — skills, RAG, coding rules
The Solution
- ✓ One tool for 18 source types: docs, repos, PDFs, videos, notebooks, wikis, and more
- ✓ Deep code analysis detects patterns, architecture, and design decisions across 27+ languages
- ✓ 12+ output platforms: Claude, Gemini, OpenAI, Kimi, DeepSeek, LangChain, Cursor, and more
- ✓ 15-45 minutes end-to-end: from any source to production-ready AI skill
10x Faster Development
Stop copying docs manually. Generate comprehensive skills in minutes, not hours.
Framework Expertise
Give AI assistants deep knowledge of any framework with API references and examples.
Always Up-to-Date
Re-run when docs update. Keep your AI knowledge fresh and accurate.
Complete AI Knowledge Toolkit
Transform any of 18 source types into structured AI skills and RAG knowledge
18 Source Types
Docs, GitHub, PDF, video, Word, EPUB, Jupyter, OpenAPI, AsciiDoc, PPTX, HTML, RSS, man pages, Confluence, Notion, Slack/Discord, and local codebases
Agent-Agnostic Architecture
All enhancers support Claude, Kimi, Codex, Copilot, OpenCode, and custom agents via unified AgentClient
Three-Stream Analysis
Split GitHub repos into Code (C3.x), Docs, and Insights streams for comprehensive skills
Smart SPA Discovery
Three-layer discovery engine: sitemap.xml, llms.txt, and SPA nav rendering for JavaScript sites
40 MCP Tools
AI agents can prepare their own knowledge with 40 MCP tools across 10 categories
Marketplace Publisher
Publish skills to Claude Code plugin marketplace repos with MarketplacePublisher and ConfigPublisher
Browser Rendering
Render JavaScript SPA sites with Playwright — auto-installs Chromium, handles React/Vue/Angular
Doctor Command
8 diagnostic checks: Python version, dependencies, API keys, MCP server, output directory
Prompt Injection Detection
Bundled security workflow scans scraped content for injection patterns and flags suspicious content
12 LLM Platform Targets
Claude, Gemini, OpenAI, Kimi, DeepSeek, Qwen, OpenRouter, Together, Fireworks, MiniMax, OpenCode, Markdown
18 Agent Install Paths
Install skills to Claude, Cursor, Windsurf, Cline, Continue, Roo, Aider, Bolt, Kilo, Kimi, and more
Video Scraping
Extract code, transcripts, and structured knowledge from YouTube and local videos with OCR
+12 more features
Ready to transform your documentation?
Get Started NowGet Started in 3 Steps
From zero to production-ready skill in 15-45 minutes
1. Install
Install from PyPI in seconds
pip install skill-seekers2. Create Skill
From any of 18 source types
skill-seekers create https://react.dev
# Or: skill-seekers create facebook/react
# Or: skill-seekers create ./my-project3. Package & Deploy
Export to any AI platform
skill-seekers package output/react --target claude
# Or: --target langchain, gemini, openai, cursor, ...Multiple Installation Options
PyPI (Recommended)
Easiestpip install skill-seekers uv (Modern)
Fastuv tool install skill-seekers From Source
Devgit clone && pip install -e . MCP Integration
18 Agents./setup_mcp.sh Who Uses Skill Seekers?
From solo developers to enterprise teams
For Developers
Create skills from documentation + GitHub repos with automatic conflict detection.
"Build a React skill from official docs + GitHub repo, catch API changes before they surprise you."
For Game Developers
Generate comprehensive skills for game engines like Godot (handles 40K+ pages!).
"Create complete Godot skill covering all topics with intelligent router/hub pattern."
For Teams
Combine internal docs + code repos + Confluence wikis into single source of truth.
"Share custom configs via private git repos or publish to the marketplace."
For Learners
Build comprehensive skills from docs, code examples, videos, and PDF tutorials.
"Combine official docs + GitHub examples + YouTube tutorials + PDF manual into one learning resource."
For Open Source
Analyze repos to find documentation gaps and outdated examples automatically.
"Detect discrepancies between documentation and actual code implementation."
Multi-Platform Support
Export your skills to 12+ LLM platforms with platform-specific optimizations
By the Numbers
Trusted metrics from a production-ready tool
Open source • MIT Licensed • Active development