Documentation Catalog#
Browse all indexed documentation artifacts organized by source. Each artifact is a reference card pointing to an authoritative external doc page, enriched with metadata, topic tags, and editorial notes.
OpenAI#
- API Platform — GPT models, function calling, embeddings, reasoning
- Agents SDK — Multi-agent orchestration with handoffs
- Symphony — Autonomous work management: issue tracker to PR
Anthropic#
- Platform — Claude API, tool use, prompt engineering, Agent SDK
- Claude Code — Agentic CLI tool for software engineering
- MCP — Model Context Protocol for agent-to-tool communication
Google#
- ADK — Agent Development Kit with workflow agents and graph orchestration
Microsoft#
- Agent Framework — Enterprise agent orchestration (successor to AutoGen)
- AutoGen — GroupChat multi-agent coordination (deprecated)
Other Providers#
- Mistral AI — Open-weight models, API, agents, fine-tuning
- Cohere — Enterprise LLMs, RAG, embeddings, reranking
- DeepSeek — Open-source reasoning models, cost-effective inference
- Together AI — Open-source model hosting, inference, fine-tuning
- Fireworks AI — Fast inference and fine-tuning for open models
Protocols & Standards#
- A2A (Agent-to-Agent) — Cross-platform agent communication (Google / Linux Foundation)
- AG-UI (Agent-User Interaction) — Agent-to-frontend communication (CopilotKit)
Methodologies#
- Spec-Driven Development — SDD methodology, AI-assisted workflows, and specification tooling
Agent Frameworks#
- LangChain — Composable LLM framework (includes LangGraph)
- CrewAI — Role-based multi-agent orchestration
- DSPy — Programming (not prompting) language models
- Pydantic AI — Type-safe agent framework with Pydantic validation
- Instructor — Structured output extraction from LLMs
Infrastructure & Evaluation#
- Pinecone — Vector database for similarity search
- Chroma — Open-source embedding database for RAG
- RAGAS — Evaluation framework for RAG pipelines
- DeepEval — Open-source LLM evaluation framework (pytest-style)
- Weights & Biases — ML experiment tracking and AI observability
- Safetensors — Secure model weight format (Hugging Face)
- PEFT — Parameter-efficient fine-tuning: LoRA, QLoRA (Hugging Face)
- TRL — Post-training: SFT, DPO, PPO, GRPO (Hugging Face)