Best Memory AI Skills & MCP Servers

Persistent memory and agent lifecycle for Claude Code — because sessions shouldn't start from zero.

Versioned context store with hybrid search for AI agents. Stdio + HTTP MCP server and CLI.

MCP server for iOS leak hunting and performance investigation. 42 tools across .memgraph, .trace, and .mxdiagnostic files. v1.18 adds analyzeMetricKitPayload (42nd tool, post-mortem production-diagnostics from MetricKit JSON payloads — crash clusters, han

AI copilot for TouchDesigner with live MCP control — 106 tools, focus + locations, hint injection, component notes, POPx inspection, knowledge corpus, project lifecycle control, technique memory, typed patch sessions, agent activity log, one-tool self-upd

Persistent key-value memory — store and recall information across sessions as JSON files.

In-memory file cache — avoid redundant disk reads with mtime-based invalidation.

Agent Model Layer — One-line LLM access with built-in memory. 29 providers, 18 embeddings, zero lock-in.

Project-local RAG memory MCP server — knowledge graph + multilingual vector + FTS5 in a single SQLite file. Per-project isolation, 30 MCP tools, codepoint-safe chunking (Korean/CJK/emoji).

Production payment infrastructure for AI agents — real Stripe/Paystack/Lightning rails, autonomous shopping with escrow, Agent Credit Score (300-850), Ed25519 identity, behavioral finance, Merkle integrity, anomaly detection, human-in-the-loop approval. B

Ricord — persistent memory + repo knowledge graph for AI coding assistants. Ships an MCP server (`ricord-mcp`) and a repo ingestion CLI (`ricord init`).

Portable project memory context CLI — bootstraps semantic enrichment workflows for any AI coding agent.

ContextStream MCP server - v0.4.x with consolidated domain tools (~11 tools, ~75% token reduction). Code context, memory, search, and AI tools.

Local-first, LLM-agnostic memory layer for AI assistants. Run via npx @psraghuveer/memento.

A Model Context Protocol server that acts as an intelligent conversation state manager and development guide for LLMs, featuring comprehensive long-term memory with persistent project artifacts

Capybara-tier CLI router — Claude Opus 4.7 with Adaptive Thinking, Strict Write Discipline, and Self-Healing Memory

Lightweight local memory layer for AI assistants — MCP server + HTTP API + CLI, backed by SQLite + FTS5. Works with Claude, Cursor, and Codex.

MCP server for Obsidian-based second brain memory using PARA methodology

MCP server that gives Claude Code persistent memory of your codebase, backed by Gemini's 2M-token context. Includes eager `ask`/`code` for cached repo Q&A and `ask_agentic` for agentic file access on repos that exceed the model's input-token limit.

MergeMind MCP server — read-only доступ к общей памяти команды агентов (markdown vault с гибридным поиском BM25 + vector + rerank). Запускается через `npx -y @agfpd/mergemind-mcp@latest` из .mcp.json плагина MergeMind (gh-репо agfpd/MergeMind).

Provider-neutral agent runtime for coding workflows. Verified DAG execution with evidence gates, worktree isolation, and local graph memory.

MCP server for SharedMemory — persistent AI memory for any tool

PLUR MCP server — persistent memory tools for any AI agent

Walrus Memory MCP client — single-binary stdio MCP server that bridges Cursor / Claude Desktop / Antigravity / Claude Code to the Walrus Memory relayer. Handles browser-based wallet login on first run.

A Model Context Protocol server for knowledge graphs designed around neurodivergent thinking patterns