Production AI Agent Stacks
Stop guessing. Find real, verified combinations of frameworks, memory, RAG, MCP tools, eval/tracing, and models that builders are actually shipping.
All Stacks
No stacks match your search. Try different keywords or filters.
How We Document Stacks
Every stack entry covers the complete production architecture, not just the model.
Orchestration
The framework or system coordinating agent behavior—LangGraph, CrewAI, OpenAI Agents SDK, etc.
Memory
How the agent maintains state—session memory, persistent storage, cross-conversation context.
RAG
Retrieval-Augmented Generation setup—vector stores, embedding models, retrieval strategies.
MCP / Tools
External integrations and capabilities—MCP servers, function calling, APIs, databases.
Eval / Tracing
How teams monitor and improve—LangSmith, Arize Phoenix, custom evaluation pipelines.
Models
The LLMs powering the agent—Claude, GPT-4, Gemini, open-source models like Llama.