03Skills · How I build

How I build

Not a list of logos — a system. Each layer calls the one below it. Select a component to see where it runs in real work.

L1 · 5 parts
Interface
Where people meet the system
L2 · 6 parts
Intelligence
Models, retrieval, and reasoning
L2 · Intelligence

RAG

Retrieval pipelines that ground model answers in source data — e.g. code embeddings in Pinecone.

Where it's used
  • RepoMind
  • This site
Open RepoMind
L3 · 5 parts
Services
APIs, streaming, orchestration
L4 · 4 parts
Data
State, memory, and storage
L5 · 4 parts
Infrastructure
Shipping reliably
Full toolkit43 tools · and counting
  • OpenAI
  • Anthropic Claude
  • Google Gemini
  • LangChain
  • LangGraph
  • RAG
  • AI Agents
  • n8n
  • Pinecone
  • ChromaDB
  • pgvector
  • OpenAI Embeddings
  • Semantic Search
  • TypeScript
  • JavaScript
  • Python
  • Java
  • C++
  • SQL
  • React.js
  • Next.js
  • Tailwind CSS
  • Flutter
  • HTML
  • CSS
  • Node.js
  • Express.js
  • FastAPI
  • REST APIs
  • WebSockets
  • SSE
  • JWT
  • PostgreSQL
  • MongoDB
  • MySQL
  • Redis
  • SQLite
  • AWSEC2 · S3 · Lambda · RDS · Bedrock
  • Vercel
  • Docker
  • CI/CD
  • Git
  • GitHub

Pick a group to light it up