A curated list of awesome open-source libraries for context engineering (Long-term memory, MCP: Model Context Protocol, Prompt/RAG Compression, Multi-Agent)
mcp-servers (modelcontextprotocol) Model Context Protocol Servers
MCP Frameworks
mcp-python-sdk (modelcontextprotocol) The official Python SDK for Model Context Protocol servers and clients
fastmcp (CEO at PrefectHQ) The fast, Pythonic way to build MCP servers and clients
fastapi_mcp (tadata-org) Expose your FastAPI endpoints as Model Context Protocol (MCP) tools, with Auth!
mcp-agent (lastmile-ai) Build effective agents using Model Context Protocol and simple workflow patterns
mcp-use (mcp-use) mcp-use is the easiest way to interact with mcp servers with custom agents
golf (golf-mcp) Production-Ready MCP Server Framework • Build, deploy & scale secure AI agent infrastructure • Includes Auth, Observability, Debugger, Telemetry & Runtime • Run real-world MCPs powering AI Agents
enrichmcp (featureform) EnrichMCP is a python framework for building data driven MCP servers
✂️ Compress Context
Prompt compression
LLMLingua (microsoft) To speed up LLMs' inference and enhance LLM's perceive of key information, compress the prompt and KV-Cache, which achieves up to 20x compression with minimal performance loss.
sammo (microsoft) A library for prompt engineering and optimization (SAMMO = Structure-aware Multi-Objective Metaprompt Optimization)
Selective_Context Compress your input to ChatGPT or other LLMs, to let them process 2x more content and save 40% memory and GPU time.
500xCompressor 500xCompressor: Generalized Prompt Compression for Large Language Models
RAG compression
xRAG xRAG: Extreme Context Compression for Retrieval-augmented Generation with One Token
recomp RECOMP: Improving Retrieval-Augmented LMs with Compression and Selective Augmentation.
CompAct (dmis-lab) CompAct: Compressing Retrieved Documents Actively for Question Answering
QGC (XMUDeepLIT) Retaining Key Information under High Compression Rates: Query-Guided Compressor for LLMs
📦 Isolate Context
Multi-Agent Frameworks
MetaGPT (FoundationAgents) The Multi-Agent Framework: First AI Software Company, Towards Natural Language Programming
agno (agno-agi) Full-stack framework for building Multi-Agent Systems with memory, knowledge and reasoning.
camel (camel-ai) CAMEL: The first and the best multi-agent framework. Finding the Scaling Law of Agents.
agent-squad (awslabs) Flexible and powerful framework for managing multiple AI agents and handling complex conversations
PraisonAI (MervinPraison) PraisonAI is a production-ready Multi AI Agents framework, designed to create AI Agents to automate and solve problems ranging from simple tasks to complex challenges.
langroid (langroid) Harness LLMs with Multi-Agent Programming
LazyLLM (LazyAGI) Easiest and laziest way for building multi-agent LLMs applications.
Image source: https://blog.langchain.com/context-engineering-for-agents/Memory for AI Agents; Announcing OpenMemory MCP - local and secure memory management.
Letta (formerly MemGPT) is the stateful agents framework with memory, reasoning, and context management.
Build Real-Time Knowledge Graphs for AI Agents
Memory for AI Agents in 5 lines of code
The Open Source Memory Layer For Autonomous Agents
Profile-Based Long-Term Memory for AI Applications. Memobase handles user profiles, memory events, and evolving context
A-MEM: Agentic Memory for LLM Agents
A memory operation system for personalized AI
Your personal plug and play memory layer for LLMs
A collection of MCP servers.
Model Context Protocol Servers
The official Python SDK for Model Context Protocol servers and clients
The fast, Pythonic way to build MCP servers and clients
Expose your FastAPI endpoints as Model Context Protocol (MCP) tools, with Auth!
Build effective agents using Model Context Protocol and simple workflow patterns
mcp-use is the easiest way to interact with mcp servers with custom agents
Production-Ready MCP Server Framework • Build, deploy & scale secure AI agent infrastructure • Includes Auth, Observability, Debugger, Telemetry & Runtime • Run real-world MCPs powering AI Agents
EnrichMCP is a python framework for building data driven MCP servers
To speed up LLMs' inference and enhance LLM's perceive of key information, compress the prompt and KV-Cache, which achieves up to 20x compression with minimal performance loss.
A library for prompt engineering and optimization (SAMMO = Structure-aware Multi-Objective Metaprompt Optimization)
Compress your input to ChatGPT or other LLMs, to let them process 2x more content and save 40% memory and GPU time.
Toolkit for Prompt Compression
500xCompressor: Generalized Prompt Compression for Large Language Models
xRAG: Extreme Context Compression for Retrieval-augmented Generation with One Token
RECOMP: Improving Retrieval-Augmented LMs with Compression and Selective Augmentation.
CompAct: Compressing Retrieved Documents Actively for Question Answering
Retaining Key Information under High Compression Rates: Query-Guided Compressor for LLMs
The Multi-Agent Framework: First AI Software Company, Towards Natural Language Programming
Full-stack framework for building Multi-Agent Systems with memory, knowledge and reasoning.
CAMEL: The first and the best multi-agent framework. Finding the Scaling Law of Agents.
Flexible and powerful framework for managing multiple AI agents and handling complex conversations
PraisonAI is a production-ready Multi AI Agents framework, designed to create AI Agents to automate and solve problems ranging from simple tasks to complex challenges.
Harness LLMs with Multi-Agent Programming
Easiest and laziest way for building multi-agent LLMs applications.
