🤖 AI Agents — Knowledge Base
My go-to hub for AI agent workflows. Built on the LangChain family. Plain English.
Topics
Foundations
What an agent is and the think-act-observe loop.
Models
Claude, OpenAI, Groq, Ollama via one helper.
Tools
@tool functions and MCP integrations.
Memory
Short-term checkpointers, long-term stores.
Middleware
The 1.0 control layer: PII, summary, custom.
Observability
LangSmith traces, evals, cost.
Harnesses & Deep Agents
Scaffolding + the batteries-included harness.
Multi-agent
Supervisor and swarm patterns.
Human-in-the-loop
Interrupts and approvals for risky actions.
RAG / retrieval
Embeddings, vector stores, retriever-as-tool.
Deployment
LangGraph Platform and self-hosting.
Multimodal
Text + image (and audio/video) in one message.
Structured output
Typed JSON out, with validation and retries.
What's planned next
These topics are on the roadmap and will get their own pages as they land Planned:
- Streaming to a UI (tokens and steps)
- Context engineering (what enters, stays, leaves the window)
- Skills (SKILL.md) for Deep Agents
- Sandboxes & safe code execution
- Subgraphs, cost & rate-limit control
The full list with status lives in the guide's master topic list.