Local persistent memory store for LLM applications including claude desktop, github copilot, codex, antigravity, etc.

For Claude.

Kcode lets you run long, tool heavy coding sessions without blowing up token costs by compressing old context into references and only restoring exact data when needed, reducing hallucinations by grounding the model in real, retrievable source data instead of guesswork.

MCP Context Server — a FastMCP-based server providing persistent multimodal context storage for LLM agents.

Attention-native memory retrieval for AI agents, using reusable KV cache to retrieve evidence from million-token memory with SOTA results on memory and code search benchmarks.

Durable hybrid memory for AI agents: vector + BM25 + temporal + graph recall, exposed through MCP, HTTP, and Python.

把单个 AI 助手升级为可扩展的多专家协作运行时， 带工业级记忆检索、会话管理和完整的开发者可观测性。

Provide precise OpenClaw documentation and exact configs to your coding agent for accurate answers without guesswork or summaries.

Minimal starter for hybrid memory retrieval in OpenClaw-style agent workflows.

Memory Utility Networks (MUN): A research-oriented framework for utility-based memory retrieval and example selection in long-horizon decision making and in-context learning. Includes reproducible experimentation, validation audits, and ablation studies investigating learned memory utility.

Replication of Recanatesi et al. (2015) https://www.frontiersin.org/articles/10.3389/fncom.2015.00149/full

Enable persistent memory for Claude that works across crashes and restarts in both Claude Code and Claude Cowork environments.

Scripts used to analyze behavioral responses during memory retrieval and the association between fMRI-BOLD activation during memory encoding and behavioral responses during memory retrieval described in Ruiz-Rizzo et al. (2022), Archives of Cinical Neuropsychology

📝 Create and manage your notes with Memos, the open-source, self-hosted service that ensures your data stays private and in your control.

Local-first AI memory store with synaptic dynamics (strengthen/decay/prune) and neighbor-based pattern expansion (L2).