Agent frameworks & infrastructure

Agent memory & context

4,032 agents · most starred first

Agent memory & context, page 38 of 41

swarmauri_similarity_gzip

Gzip-based compression similarity for Swarmauri.

synapto-llm

Synaptic Weight Eviction engine for dynamic LLM memory consolidation

true-memory-fragments

Stale-context protection and source-aware code memory for AI coding agents

docmancer

Docmancer review: compare features, pricing, use cases, access model, and alternatives for this Memory agent in 2026. Local AI agent memory tool with MCP server and plugins

@3ngram/server

MCP server and REST API for persistent, typed AI-agent memory.

delx-memory

Local-first multi-agent memory MCP. Shared SQLite + FTS5, secret-blocking, lite transport (no SDK by default), namespaces, batch + handoff tools.

session-atlas

Self-hosted multi-device, multi-agent session atlas

isotope-zero

Sub-millisecond, local-first cognitive memory layer for AI agents — NumPy/BLAS-accelerated hybrid retrieval, Ebbinghaus decay, and knowledge compaction. Ships a premium React/Ink terminal UI plus the `izero` Python CLI bridge.

crewai-persistence-dynamodb

Amazon DynamoDB persistence backend for CrewAI Flows with built-in message pruning via agentstate-reducer

dma-langgraph

Thin LangGraph node helpers for DMA memory.

langgraph-agentram

AgentRAM-backed long-term memory for LangGraph and LangChain: a BaseStore plus agent-callable memory tools, no vector database.

memstrata-langchain-conversational

LangChain retriever integration for MemStrata Conversational Memory

pydantic-ai-cosmosdb-persistence

Azure CosmosDB backend for pydantic-ai-persistence — StepStore + message history for PydanticAI

pydantic-ai-dynamodb-persistence

Amazon DynamoDB backend for pydantic-ai-persistence — StepStore + message history for PydanticAI

pydantic-ai-firestore-persistence

Google Firestore backend for pydantic-ai-persistence — StepStore + message history for PydanticAI

pydantic-ai-persistence

Backend-agnostic persistence for PydanticAI — chat history + durable StepStore, over a small async KV interface

shared-mdstorage-client

Typed sync and async client for the Shared MDStorage HTTP contract

shelfmark-rag

Local hybrid retrieval (BM25 + embeddings + RRF) for agent memory and codebases — one SQLite file, MCP server included.

strands-github-storage

GitHub-repository Storage backend for Strands Agents — persists keys as versioned, diffable files in a repo.

strands-mongodb-storage

MongoDB implementation of the Strands Agents Storage interface — durable bytes-under-keys for sessions, context offload, and memory

strands-sql-storage

SQL (SQLAlchemy) implementation of the Strands Agents Storage interface — durable bytes-under-keys for sessions, context offload, and memory

strands-storage-dynamodb

Amazon DynamoDB implementation of the Strands Agents Storage interface — durable bytes-under-keys for sessions, context offload, and memory

agent-history

Local CLI for finding and resuming AI agent sessions across local projects.

coral-memory

脑珊瑚 Coral Memory — 面向 LLM Agent 的热度感知记忆层(DeepSeek Harness 插件):MCP 记忆工具(memory_search/insert/thread_* 等 16 个)+ 设置面板(统计/容量/检索/配额)。作者:Mr. Code Muggle (@Ne)。

@deepseek-ai/dsh-compaction

Abstract compaction service seam (ctx.compaction) for the DeepSeek Harness

@deepseek-ai/dsh-compaction-basic

Token-meter-driven compaction policy and LLM summarization backend for the DeepSeek Harness

@deepseek-ai/dsh-session

Event-sourced session store for the DeepSeek Harness

@deepseek-ai/dsh-session-persistence-jsonl

JSONL durable session persistence backend for the DeepSeek Harness

@deepseek-ai/dsh-session-reference

Cross-session snapshot references and durable untrusted model context (ctx.sessionReferences)

@nestjs-agentic/memory

Multi-tier cognitive memory module for nestjs-agentic enabling Short-Term, Scratchpad, and Composite memory stores.

@hybridlabor-api/memb

memB - Local Offline Long-Term Agentic Memory

ise-harness

Coding Agent Harness SDK — 记忆与上下文管理

@langgraph-toolkit/adapter-checkpointers

Database checkpointers for @langgraph-toolkit/core: SQL (SQLite/Postgres/MySQL), Redis, MongoDB. Drivers are injected, so no database library is a hard dependency.

agent-experience

Event-sourced experience capture, validation, and reuse for AI agents.

agents-memory

Local markdown memory & cross-agent context engine for coding agents (MCP).

areev-crewai

Areev storage backend for CrewAI memory: supersession-based updates, audited deletes, and right-to-erasure over memory sources

areev-langgraph

Areev persistence for LangGraph: a BaseStore and a BaseCheckpointSaver over content-addressed, governable memory files

citadeldb-ms-agent-framework

Microsoft Agent Framework chat history backed by Citadel: encrypted at rest, with deletes that destroy the key

citadeldb-openai-agents

OpenAI Agents SDK sessions backed by Citadel: encrypted at rest, with deletes that destroy the key

citadeldb-strands-agents

Strands Agents session storage backed by Citadel: encrypted at rest, with deletes that destroy the key

cortadel-crewai

Cortadel memory for CrewAI — self-hosted long-term temporal graph memory as a drop-in crewai.Memory, plus native tools.

cortadel-pydantic-ai

Cortadel long-term graph memory as a native Pydantic AI capability — memory tools plus automatic recall and persistence.

crewai-persistence-mongodb

MongoDB persistence backend for CrewAI Flows with built-in message pruning via agentstate-reducer

crewai-persistence-sql

SQL persistence backend for CrewAI Flows via SQLAlchemy (SQLite, PostgreSQL, MySQL) with built-in message pruning via agentstate-reducer

langgraph-dynamodb-store

A DynamoDB-backed BaseStore for LangGraph — cross-thread memory for agents, no Postgres/pgvector required

llama-index-memory-itsuki

Itsuki long-term memory block for LlamaIndex.

llmwiki-harness

Context-Memory Harness for AI Agents. Context Window = RAM, Local Wiki = Disk.

nowledge-mem-langgraph

Identity-aware Nowledge Mem connector for LangGraph and LangChain agents

strands-postgres-storage

PostgreSQL implementation of the Strands Agents Storage interface — durable bytes-under-keys for sessions, context offload, and memory

strands-storage-mongodb

MongoDB implementation of the Strands Agents Storage interface — durable bytes-under-keys for sessions, context offload, and memory (aka strands-mongodb-storage)

strands-storage-postgres

PostgreSQL implementation of the Strands Agents Storage interface — durable bytes-under-keys for sessions, context offload, and memory (aka strands-postgres-storage)

three-layer-agent-memory

A filesystem-native, version-controllable long-memory paradigm for AI agents. Surface/Middle/Deep layers + recall/writeback/consolidate protocol. Pure Markdown, zero infra, drives agent self-evolution.

tracegc-crewai

CrewAI adapter for tracegc context compaction

tracegc-langgraph

LangGraph adapters for tracegc

tracegc-storage

Storage adapters for tracegc

xyberos-storage

Object storage plugin (RFC-0019, M9): S3, Azure Blob, GCS, OneDrive behind one ObjectStore contract

@reallexi/rmai-ui-linux-amd64

Terminal UI binary for the Real Memory AI Agent CLI (linux/x64).

@reallexi/rmai-ui-windows-amd64

Terminal UI binary for the Real Memory AI Agent CLI (win32/x64).

@reallexi/rmai-ui-linux-arm64

Terminal UI binary for the Real Memory AI Agent CLI (linux/arm64).

@reallexi/rmai-ui-darwin-arm64

Terminal UI binary for the Real Memory AI Agent CLI (darwin/arm64).

@reallexi/rmai-ui-windows-arm64

Terminal UI binary for the Real Memory AI Agent CLI (win32/arm64).

llm-wiki-loop

The production framework for self-improving & self-organizing LLM knowledge vaults — grounding invariants, code drift detection, auto-skillification, and 1-click multi-agent setup.

@vectorize-io/hindsight-eve

Automatic long-term memory for Vercel Eve agents — Hindsight memory injected before each turn and retained after, with no model tool-calling

mottainai

Coding-agent orchestration and MCP context runtime with governed repository execution, bounded tool results, and multi-agent session management.

@nestjs-agentic/experience

Experience learning, trajectory reflection, and self-correction engine for nestjs-agentic.

@cortadel/langgraph

Cortadel long-term memory for LangGraph — a BaseStore implementation, memory tools, and automatic recall/persist graph nodes.

claude-contextmesh

Intelligent context layer for Claude Code — infinite session memory, minimal active context

crewai-memorysync

MemorySync memory for CrewAI: a native StorageBackend for Memory(storage=...), agent memory tools, and recall context for tasks.

langchain-memorysync

MemorySync integration for LangChain: chat message history, memory tools, and recall context.

openai-agents-memorysync

MemorySync memory for the OpenAI Agents SDK: a drop-in Session implementation with durable server-side history, long-term memory instructions, and agent memory tools.

@tages/cli

CLI for Tages AI agent memory

getmnemo-langgraph

Mnemo-backed long-term memory for LangGraph JS — a BaseStore implementation plus remember/recall tools.

@inupedia/spotlight-memory

Optional exact and semantic answer-cache primitives for Inupedia Spotlight

aegisdb-langgraph

Use AegisDB as a LangGraph long-term-memory store

deep-memory-agent

Factory for building LangChain deepagents equipped with episodic, semantic, and procedural memory.

khwan-crewai

Khwan memory for CrewAI — recall/remember/verify as tools, and a prepare→record wrapper around a crew run. Khwan never runs your model.

langchain-mongodb-deepagents-vfs

MongoDB Atlas-backed virtual filesystem search adapter for LangChain DeepAgents

nidra-agent-memory

The sleep cycle for AI memory: every memory carries its evidence; a scheduled pass keeps the evidence honest.

@awareness.market/local

Local-first AI agent memory system. No account needed.

cadet-token-saver

Cadet Token Saver — an orchestration + measurement layer that reduces AI-agent context consumption.

@reallexi/rmai-ui-darwin-amd64

Terminal UI binary for the Real Memory AI Agent CLI (darwin/x64).

@i-xor/lwc

Agent-driven proactive memory CLI for AI agents — autonomously recall, maintain, and evolve persistent, source-grounded knowledge across sessions.

memorysync-langchain

MemorySync integration for LangChain.js and LangGraph.js: chat message history, memory tools, recall context, and a LangGraph store.

pi-active-memory

Proactive semantic memory and mid-run steering for the Pi coding agent

dsh-sgme

SGME memory engine bridge for DeepSeek Harness — multi-agent shared long-term memory via HTTP.

@shibenenen/mem

Local-first Rust memory service for multi-agent workflows — prebuilt-binary installer.

@rsl-innovation/mem-port

A portable, long-term agentic memory MCP server — a thumb drive for your AI context.

@finchagentic/mcp

The runtime layer for Agentic AI. Persistent memory, autonomous agents, and workflows that survive every session.

@matthewfl/pi-contemplator

A Pi extension that keeps long-running agentic sessions on track with background memory, contemplation, and structural review.

agent-chat-archiver

Archive local AI chat histories to S3, freeing up disk space while keeping conversations searchable

agent-filetree-memory-mcp

Encrypted, capability-scoped, versioned Markdown file-tree memory for MCP agents

agent-framework-memorysync

MemorySync for Microsoft Agent Framework: a ContextProvider with budgeted instruction-layer recall, per-run identity, once-per-turn duplicate-proof capture, and fail-open everything.

agent-memory-bridge

Governed project memory for AI coding agents across sessions and tools

ai-agent-persistence

Evidence-first tools for validating AI agent persistence publications and probing durable state serialization

autogen-memorysync

MemorySync for Microsoft AutoGen: an async-native Memory implementation with a hard recall budget, role-aware retrieval, duplicate-proof fact capture, and session-scoped clear.

bedrock-memorysync

MemorySync reference architecture for AWS Bedrock agents: a memory-augmented Converse loop and a Bedrock Agents action-group Lambda — the first vendor action-group memory integration.

claude-rework

Local, automatic memory for Claude Code. Remembers past sessions in 0.7s for ~2.5k tokens instead of 211k. Nothing leaves your machine.

cnb-agentic-memory

CNB Issue 智能体记忆系统(cnb-agentic-memory)是基于 CNB 平台构建的通用记忆工具,把 CNB 的 issue、知识库、检索能力组合成一个开箱即用的智能体记忆层。

context-m-langchain

LangChain BaseMemory adapter for Context-M — the Universal Neuro-Symbolic Memory Fabric.

continuum-agent

Verifiable semantic recovery layer for long-running AI agents.

More in Agent frameworks & infrastructure