MCP CheckMCP servers for memory
Every memory server in the official MCP Registry, 729 in all, ranked by how closely it fits. 572 work without an API key and 1 come from the company itself. Each row shows the security checks, the secrets asked for and an install command.
memory MCP servers 729
Now showing are unlisted. These entries didn't pass our listing filter: no source repo, a very short or duplicated description, or a publisher with 100+ servers. Same automated checks, shown as-is. How we check
| Type | Secrets | Version | Actions | ||||
|---|---|---|---|---|---|---|---|
★ 48k io.github.DeusData/codebase-memory-mcp Codebase knowledge graph for AI agents — 162 languages, sub-ms queries, 99% fewer tokens. | 6/71 warning | 76health of 100 | npmRuns locally | Noneno keys asked | 0.12.1 | ||
★ 4.4k io.github.basicmachines-co/basic-memory Local-first knowledge management with bi-directional LLM sync via Markdown files. | 6/71 warning | 88health of 100 | PyPIRuns locally | Noneno keys asked | 0.24.1 | ||
★ 1.3k io.github.chunxiaoxx/nautilus-compass Local-first long-term memory for AI agents, with byte-recomputable signed verification receipts. | 7/81 warning | 62health of 100 | PyPILocal or hosted | Noneno keys asked | 3.2.0 | ||
 ★ 989 io.github.neo4j-contrib/mcp-neo4j-memory Official Neo4j MCP Neo4j Knowledge Graph Memory Server | 7/7All passed | 75health of 100 | PyPIRuns locally | 1secret asked | 1.0.0 | ||
★ 333 io.github.aayoawoyemi/ori-memory Persistent memory infrastructure for AI agents. Identity, knowledge graph, and search. | 7/7All passed | 76health of 100 | npmRuns locally | Noneno keys asked | 0.3.0 | ||
 ★ 316 io.github.varun29ankuS/shodh-memory Cognitive memory for AI agents — semantic search, Hebbian learning, knowledge graphs. | 6/71 warning | 74health of 100 | npmRuns locally | 1secret asked | 0.2.0 | ||
★ 270 io.github.sachitrafa/cognitive-ai-memory Ebbinghaus-based persistent memory for Claude. Memories decay with time, strengthen on recall. | 4/4All passed | 51health of 100 | DockerRuns locally | Noneno keys asked | 1.0.2 | ||
★ 266 io.github.ZipengWu365/tomc-memory CPU task-specific context preparation and explicitly saved local project notes through MCP. | 5/61 warning | 50health of 100 | MCPBRuns locally | Noneno keys asked | 0.1.1 | ||
★ 242 io.github.nhadaututtheky/neural-memory Persistent memory for AI agents — 55 MCP tools, spreading activation recall. | 4/4All passed | 80health of 100 | npmRuns locally | Noneno keys asked | 4.27.2 | ||
★ 224 io.github.omega-memory/omega-memory Persistent memory, coordination, and learning for AI agents. Local-first, 25 MCP tools. | 4/4All passed | 82health of 100 | PyPIRuns locally | Noneno keys asked | 1.3.0 | ||
★ 197 io.github.modusensus/mneme-memory Cross-session memory for AI coding agents via six tools; local-first SQLite store. | 7/7All passed | 76health of 100 | npmRuns locally | Noneno keys asked | 0.1.1 | ||
★ 100 io.github.cyberlife-coder/velesdb-memory Offline agentic memory: remember/recall/relate/forget/why over a fused vector+graph+columnar engine | 6/6All passed | 68health of 100 | MCPBRuns locally | Noneno keys asked | 0.15.0 | ||
★ 100 io.github.yuvalsuede/memory-mcp Persistent memory + git snapshots for Claude Code. Never lose context or code. | 7/7All passed | 49health of 100 | npmRuns locally | Noneno keys asked | 1.2.1 | ||
★ 77 ai.maximem/synap Persistent memory for AI agents — log and recall conversation context over MCP. | 4/4All passed | 60health of 100 | RemoteHosted | Noneno keys asked | 0.1.0 | ||
 ★ 76 io.github.Vortx-AI/emem Shared, verifiable memory for machines and AI. Encode where data lives; decode with any AI. | 5/61 warning | 53health of 100 | DockerLocal or hosted | Noneno keys asked | 2.4.4 | ||
★ 73 io.github.vbcherepanov/total-agent-memory Persistent local memory for coding agents: temporal knowledge graph, procedural and episodic recall | 6/71 warning | 56health of 100 | PyPIRuns locally | Noneno keys asked | 14.9.0 | ||
★ 65 io.github.MihaiBuilds/memory-vault Local-first AI memory layer with hybrid search. Postgres + pgvector. Self-hosted, MIT. | 4/4All passed | 69health of 100 | DockerRuns locally | 1secret asked | 1.6.1 | ||
★ 56 io.github.alphaonedev/ai-memory Persistent memory for any AI — zero token cost until recall | 4/4All passed | 65health of 100 | DockerRuns locally | Noneno keys asked | 0.5.1 | ||
★ 49 io.github.kapillamba4/code-memory Local semantic code search with Git history. Works offline, no API key needed. | 7/7All passed | 52health of 100 | PyPIRuns locally | Noneno keys asked | 1.0.33 | ||
★ 44 io.github.ipiton/agent-memory-mcp Memory, docs, and repo context layer for engineering agents. | 2/2All passed | 60health of 100 | RemoteRuns locally | Noneno keys asked | 0.13.4 | ||
★ 43 io.github.Wuesteon/lean-memory Embedded, local-first agent memory: one SQLite file per namespace, no server, no cloud. | 7/7All passed | 61health of 100 | PyPIRuns locally | Noneno keys asked | 0.3.0 | ||
★ 40 io.github.putervision/state-memory-mcp Persistent workflow graph server tracking tasks, decisions, spatial entities, and blockers. | 7/7All passed | 68health of 100 | npmRuns locally | Noneno keys asked | 1.4.1 | ||
★ 39 io.github.putervision/vision-memory-mcp Local visual UI cache for AI agents using perceptual hashing, CLIP, and 3D spatial grounding. | 7/7All passed | 68health of 100 | npmRuns locally | Noneno keys asked | 1.4.2 | ||
★ 37 io.github.gonzaroman/ihmt-memory Long-term memory for AI coding agents, stored as a tree of plain local files. | 7/7All passed | 66health of 100 | PyPIRuns locally | Noneno keys asked | 1.2.1 | ||
★ 35 io.github.gambletan/cortex Local-first AI memory engine — 4-tier memory, people graph, Bayesian beliefs. Encrypted, 62µs. | 4/4All passed | 64health of 100 | DockerRuns locally | Noneno keys asked | 1.8.0 | ||
★ 31 ink.cairn/memory Cross-session memory for AI agents with a Source Receipt for every memory, over MCP. | 3/3All passed | 60health of 100 | RemoteHosted | Noneno keys asked | 0.1.0 | ||
★ 31 io.github.Astrix-Labs/genesys-memory Open-source causal memory for AI agents: persistent, explainable, MCP-native memory (13 tools). | 7/7All passed | 56health of 100 | PyPIRuns locally | 2secrets asked | 0.5.2 | ||
★ 30 io.github.skynetcmd/m3-memory Local-first, cloud-friendly memory — 100+ tools, 99.2% LongMemEval-S retrieval@8, hybrid search. | 6/71 warning | 62health of 100 | PyPIRuns locally | Noneno keys asked | 2026.10.25.0 | ||
★ 30 io.github.LeandroPG19/memory-industry MemoryIndustry persistent memory MCP server: BM25+MMR retrieval and audit. | 6/71 warning | 56health of 100 | npmRuns locally | Noneno keys asked | 0.30.0 | ||
★ 29 io.github.IgorGanapolsky/mcp-memory-gateway Pre-action gates that block AI agents from repeating known mistakes. | 6/71 warning | 71health of 100 | npmRuns locally | Noneno keys asked | 0.7.4 | ||
★ 27 io.github.mnemoverse/mcp-memory-server Hosted agent memory that learns from outcomes, with shared rooms. No key: mcp.mnemoverse.com/mcp | 8/8All passed | 66health of 100 | npmLocal or hosted | 1secret asked | 0.16.0 | ||
★ 27 ai.zensation/zenbrain Agent memory: working, episodic, semantic, procedural and core layers. Local SQLite, no account. | 7/7All passed | 61health of 100 | npmRuns locally | Noneno keys asked | 0.3.0 | ||
★ 24 io.github.Thanatos9404/llmslim Long-term memory and context fitting for agents, with no LLM in the loop. Local SQLite. | 7/7All passed | 60health of 100 | PyPIRuns locally | Noneno keys asked | 0.11.2 | ||
★ 24 io.github.hermes-labs-ai/fidelis-memory Recall locally stored agent memories through an MCP-compatible host. | 7/7All passed | 67health of 100 | PyPIRuns locally | Noneno keys asked | 0.2.0 | ||
★ 24 io.github.jmeyer1980/neurodivergent-memory Persistent knowledge graph MCP server for neurodivergent thinking. BM25 search, no cloud LLM. | 7/7All passed | 67health of 100 | npmRuns locally | Noneno keys asked | 0.1.8 | ||
★ 24 io.github.sgx-labs/same Persistent memory for AI coding agents. Local-first SQLite + vector search. 17 MCP tools. | 6/71 warning | 57health of 100 | npmRuns locally | Noneno keys asked | 0.12.1 | ||
★ 24 io.github.Bumblebiber/hmem-mcp Persistent 5-level hierarchical memory for AI agents. SQLite-backed, lazy-loaded. | 7/7All passed | 58health of 100 | npmRuns locally | Noneno keys asked | 1.6.7 | ||
★ 23 io.github.MontyGovernance/montycat-mcp Shared, persistent memory for AI agents, with semantic recall and live updates. | 7/7All passed | 54health of 100 | PyPIRuns locally | 1secret asked | 1.1.6 | ||
★ 23 io.github.MontyGovernance/memocat-mcp One shared, persistent memory across AI agents and systems, with semantic recall and live updates. | 7/7All passed | 64health of 100 | PyPIRuns locally | 1secret asked | 0.4.3 | ||
★ 22 io.github.writerslogic/holographic-memory-mcp Local persistent memory for AI agents with recall, search, forgetting, and encrypted storage. | 7/7All passed | 60health of 100 | npmRuns locally | Noneno keys asked | 0.1.2 | ||
★ 22 dev.adrkit/mcp Deterministic, offline, read-only ADR decision memory for coding agents. No model or network calls. | 7/7All passed | 76health of 100 | npmRuns locally | Noneno keys asked | 0.20.0 | ||
★ 20 io.github.neverinfamous/memory-journal-mcp Developer journal with knowledge graphs, relationships, Git context, and semantic search | 4/4All passed | 49health of 100 | DockerRuns locally | Noneno keys asked | 1.1.3 | ||
★ 16 io.github.RamaAditya49/titen-memory Evidence-grounded agent memory with mandatory provenance, on local SQLite or Cloudflare D1 | 7/7All passed | 75health of 100 | npmRuns locally | 1secret asked | 0.7.2 | ||
★ 16 io.github.ebeirne/lians Bitemporal memory for AI agents with point-in-time recall and tamper-evident audit history. | 7/7All passed | 68health of 100 | PyPIRuns locally | 1secret asked | 0.5.0 | ||
★ 15 io.github.xChuCx/agent-memory Git-native project memory for AI coding agents: Markdown source of truth, reviewable, secret-safe. | 7/7All passed | 66health of 100 | npmRuns locally | Noneno keys asked | 0.6.1 | ||
★ 15 io.github.michielinksee/linksee-memory Hand a project over with the reasons attached. Local-first cross-agent memory MCP + drift detection. | 7/7All passed | 66health of 100 | npmRuns locally | Noneno keys asked | 0.15.2 | ||
★ 15 io.github.studiomeyer-io/local-memory-mcp Persistent local AI memory. SQLite + FTS5 + vector search, knowledge graph. No cloud, no API keys. | 7/7All passed | 64health of 100 | npmRuns locally | Noneno keys asked | 2.4.3 | ||
★ 15 io.github.drQedwards/pmll-memory-mcp PMLL Memory MCP — persistent KV context memory and Q-promise deduplication. | 7/7All passed | 64health of 100 | npmRuns locally | Noneno keys asked | 0.1.1 | ||
★ 14 io.github.ozankasikci/global-agent-memory Shared, project-aware memory for Claude Code, Codex, and other agents with human review. | 7/7All passed | 56health of 100 | PyPIRuns locally | Noneno keys asked | 0.1.6 | ||
★ 14 io.github.michael-denyer/hot-memory-mcp Two-tier memory: hot cache (0ms) + semantic search. Self-organizing. | 7/7All passed | 65health of 100 | PyPIRuns locally | Noneno keys asked | 0.8.0 |
Showing 1 to 50 of 729. These are the most recently updated; search or pick a category for the rest.
memory MCP servers: common questions
What is the best MCP server for memory?
mcp-neo4j-memory is published by the company behind memory, which makes it the safest first pick. On GitHub stars, Codebase Memory (★ 48k), basic-memory (★ 4.4k) and Nautilus Compass Memory (★ 1.3k) lead the 729 we list. mcp-neo4j-memory is named for memory and passed every check we run.
Can I trust these memory MCP servers?
585 of the 729 passed every registry check. Reading the package code or tool descriptions of 301 of them turned up 71 with something to look at, from install scripts to obfuscated code. A clean result means none of our patterns matched, not that a server is safe.
MCP server or Claude skill for memory?
An MCP server gives the AI new tools, such as reaching memory itself. A skill only teaches it how to do a task with tools it already has. There are 739 Claude skills for memory if that fits better.
Hosted or local: how do memory MCP servers run?
153 can run hosted, so you only paste a URL into Claude, Cursor or VS Code. The 611 that run locally ship as 287 npm, 211 PyPI, 45 Docker, 42 MCPB bundle and 4 Cargo packages.
Are memory MCP servers kept up to date?
301 of the 729 were updated in the 30 days before 31 Oct 2026. The latest was Builderforce Memory, on 28 October.
More topics: every “MCP server for…” page. Prefer a skill? Claude skills for this (739).