learn-mcp

AI Engineering Learning Reference MCP Server — 801 searchable references for 2026.

Topics

Topic	Description	Refs
mcp	Model Context Protocol — servers, clients, deployment	73
rag	Retrieval-Augmented Generation — vector DBs, chunking, reranking	60
agents	Agentic AI — LangGraph, CrewAI, AutoGen, tool use	79
fine-tuning	LLM Fine-Tuning — LoRA, QLoRA, PEFT, deployment	59
mlops	MLOps & LLMOps — CI/CD, Docker, K8s, MLflow	62
prompt-engineering	Advanced Prompting — CoT, ReAct, ToT, structured output	56
inference	LLM Inference — vLLM, Ollama, SGLang, quantization	86
claude-code	Claude Code — context engineering, hooks, skills, worktrees	58
vector-db	Vector Databases — Qdrant, Pinecone, Weaviate, Milvus	61
gpu-compute	GPU & CUDA — PyTorch, kernels, mixed precision	64
autonomous-agents	Autonomous Systems — self-correcting loops, memory, planning	64
seo	SEO & Web Promotion — landing pages, design psychology, Core Web Vitals, marketing	79

Tools

• list_topics — List all available topics
• search_references(query, topic?, limit?) — Keyword search across references
• get_topic_references(topic) — Get all references for a topic
• get_workflow(topic, workflow_name?) — Get step-by-step workflows
• get_examples(topic) — Get code examples and config snippets
• reference_stats() — Statistics about the reference database
• index_to_qdrant() — Index all references into Qdrant for semantic search
• semantic_search(query, topic?, limit?) — Vector similarity search via Qdrant

Quick Start

Docker (recommended)

docker compose up -d learn

Local Development

cd ai-stack/learn-mcp
pip install "mcp[cli]>=1.6.0" qdrant-client sentence-transformers
python server.py

Build References

python merge_refs.py  # Merges data/*.json into references.json

Architecture

learn-mcp/
  server.py          # FastMCP server (Streamable HTTP on port 8001)
  references.json     # Merged reference database (generated)
  merge_refs.py       # Merge + validate script
  data/               # Per-topic reference files
    mcp.json
    rag.json
    agents.json
    ...
  Dockerfile          # Python 3.14 + deps

Adding References

1. Create or edit data/<topic>.json with new references
2. Run python merge_refs.py to merge and validate
3. Restart the server to pick up changes

Each reference follows this schema:

{
  "topic": "mcp",
  "type": "course",
  "title": "DeepLearning.AI MCP Course",
  "description": "Build rich-context AI apps with Anthropic's MCP protocol...",
  "url": "https://deeplearning.ai/courses/mcp-build-rich-context-ai-apps-with-anthropic/",
  "tags": ["mcp", "fastmcp", "sse", "deployment"],
  "difficulty": "intermediate",
  "workflow": ["Step 1: ...", "Step 2: ..."],
  "code": "from mcp.server.fastmcp import FastMCP\n..."
}

License

MIT