thedotmack/claude-mem
Claude-Mem is a plugin for Claude Code that captures, compresses, and preserves coding session context to enable persistent memory and continuity across sessions.
Awesome AI Agents › Continual Learning Agents
zjunlp/LightMem
LightMem, from the ZJUNLP group, is a memory management framework for large language models and AI agents, published as an ICLR 2026 paper on lightweight and efficient memory-augmented generation. It supplies the storage, retrieval and update machinery an agent needs to carry information across long conversations or long-running tasks, packaged behind a small API so it can be added to an application in a few lines. The design is modular: storage engines and retrieval strategies are pluggable, and the factory layer supports both cloud APIs such as OpenAI and DeepSeek and locally hosted models through Ollama, vLLM and direct Transformers auto-loading, including DeepSeek reasoning-effort and thinking-mode configuration. The repository is broader than the original method and now hosts several memory approaches side by side, each with its own documentation page: LightMem itself; FluxMem, a connectivity-evolving framework that models memory as a heterogeneous graph; EM2Mem, an event-centric multimodal memory aimed at long-video question answering; and StructMem, an ACL 2026 hierarchical memory that preserves event-level bindings and cross-event connections. Evaluation is a first-class concern. The project ships ready-to-run reproduction scripts for the LoCoMo and LongMemEval benchmarks plus combined baselines, and a memory toolkit that benchmarks competing memory layers including Mem0, A-MEM, EverMemOS and LangMem on the same datasets; a companion repository, MemBase, extends this comparison. An MCP server is included so the memory tools can be called by MCP-compatible agents and clients, and tutorial notebooks plus a demo video cover long-context handling scenarios. The code is Python and MIT licensed. It is aimed at researchers comparing memory architectures and at developers who need personalisation and long-term recall in chatbots or agent systems without running a heavy retrieval stack.
https://github.com/zjunlp/LightMem
Also in Continual Learning Agents
vectorize-io/hindsight
Hindsight is an agent memory system designed to create smarter, learning AI agents, addressing limitations of traditional memory techniques like RAG and knowledge graphs.
langchain-ai/langchainjs
LangChain.js is a TypeScript framework for building context-aware, reasoning applications powered by large language models, offering modular components, chains, agents, and tools for production and deployment.
aden-hive/hive
Hive is an outcome-driven agent development framework designed for evolving agents with AI evaluation, human-in-the-loop, and self-improvement capabilities.
aiming-lab/MetaClaw
Proxy-based layer that turns everyday conversations with a personal agent into learning signals, injecting skills per turn and running scheduled reinforcement learning updates without a local GPU.
kayba-ai/agentic-context-engine
Agentic Context Engine (ACE) is a framework that enables AI agents to autonomously learn from their successes and failures, continuously improving their performance across various tasks without additional training data or fine-tuning.
AgentDock/AgentDock
AgentDock is an open-source framework for building sophisticated AI agents with configurable determinism, enabling reliable and creative AI applications through a node-based architecture and multi-stage workflows.
modelscope/AgentEvolver
Self-evolving agent training framework from ModelScope that combines automatic task generation, experience-guided exploration and attribution-based credit assignment into one reinforcement learning stack.