{ "source_file": "0930", "source_platform": "Dev.to", "author": "Edward Izgorodin", "title_listed": "I Opened All Thirteen Memory MCP Servers. Every Public Signal I Trusted Was Wrong.", "keywords": "", "domain": "大模型算法", "content_type": "深度技术解析", "is_paid_flag": "否", "url": "https://dev.to/izgorodin/i-opened-all-thirteen-memory-mcp-servers-every-public-signal-i-trusted-was-wrong-1i1g", "publish_date": "2026-08-31", "corpus_stage": "", "notes": "", "id": "6c7a7a238625", "collected_at": "2026-10-02T14:03:23", "title": "I Opened All Thirteen Memory MCP Servers. Every Public Signal I Trusted Was Wrong. - DEV Community", "body_markdown": "I went looking for a memory MCP server the way anyone does. Sort by stars, check the registry, read the README, pick one.\nThen I opened all thirteen that actually ship, and every one of those four signals turned out to measure something other than what I thought it measured. Not one of them is a lie. Each is a real number pointing at a real thing. The problem is that the thing is rarely the server you are about to install.\nHere is what each signal actually reports, with the cases where the gap is widest.\nThe star count often belongs to a different project\nThe largest number in this field is about 90,100 stars, and it belongs to claude-mem. That one is honest: the stars sit on the repository that ships the thing.\nTwo of the best-known entries are not.\nMem0's memory MCP server lives inside an umbrella repository carrying roughly 62,800 stars. Those stars cover the engine, the SDKs, the docs, and everything else in the monorepo. Supermemory has the same shape: about 28,800 stars on the monorepo where its MCP v2 lives, while the deprecated v1 repository sits near 1,700.\nMemori is the sharpest case. Its engine has around 15,700 stars. Its MCP wrapper, the part you would actually connect to a client, has\ntwo\n.\nNone of that makes those projects worse. Mem0 and Supermemory are serious, mature systems, and Memori's engine has earned its count. But if you shortlist by star count without asking which repository the number is attached to, you have ranked engines against wrappers against monorepos and called it a comparison.\nThe check takes ten seconds: open the repository the install command actually points at, and read the count there.\nThe registry is a form, not a review\nThe official MCP Registry looks like a filter. It is not one. Publishing to it is self-service: a maintainer fills in a form. Nobody reviews the server, tests it, or checks that it does what the description says.\nThe absence side is worse.\nThe largest project in the whole comparison, claude-mem at roughly 90,100 stars, has no registry entry at all.\nNeither does MemPalace at about 58,200. Meanwhile some of the smallest projects in the registry have almost no users.\nSo a registry entry tells you one thing: MCP-aware clients that browse the registry can find this server. That is genuinely useful, because such a client cannot find what is not listed. It is just not a quality signal, and reading it as one inverts the field.\nDownload counts measure the reference implementation, not the field\nThe most-downloaded package in this comparison is the official\nserver-memory\n, at roughly 390,000 npm downloads a month. That is more than an order of magnitude above anything else here.\nIts own README calls it \"a basic implementation.\"\nThat is not false modesty. It stores a knowledge graph in a single local JSONL file, searches by substring, and serves one user. It is actively maintained and it is a reasonable first experiment. It is also the thing people install to try MCP memory once, which is exactly why its download number looks the way it does.\nDownloads measure trial, and trial concentrates on whatever is easiest to install. Engine libraries distort the picture further:\nmem0ai\non PyPI counts far higher than anything in this list, but it is an engine, not an MCP server, and comparing the two is comparing a library to a service.\nA README is a snapshot, and sub-projects disappear\nAn earlier version of the comparison page behind this article pointed readers at OpenMemory as Mem0's local Docker path. That path was removed from the monorepo in July 2026. Mem0's documented MCP entry point today is a hosted server.\nNothing was hidden. The project moved, as projects do. But the sentence had been true when it was written, and it was still sitting there being read after it stopped being true.\nThe same failure mode is waiting in\nmemento-mcp\n. Its README describes a knowledge graph with vector embeddings and temporal decay over Neo4j, and that description is accurate. What the README does not say is that the last commit to main was\n10 May 2025\n, and npm downloads sit around 127 a month. You would be adopting abandoned code and, in this case, forking it.\nBefore adopting any of these, check the last commit date and whether the sub-project you are citing still exists in the repository. Both checks are faster than reading the feature list.\nOne of the most-searched names is not a product\nSearches for a memory MCP server called \"gmemory\" return confident results. There is no such shipped server. The term points at G-Memory, a multi-agent memory research framework on arXiv, paper 2506.07398.\nIt is a paper. It is a good paper. It is not something you can connect to a client.\nThe nearest shipped SQL-native option is Memori, from MemoriLabs, formerly GibsonAI, and that is a different project entirely, with the engine-versus-wrapper split described above.\nFree does not mean free, and free does not mean self-hosted\nMost of this field converges on roughly $15 to $30 a month for a hosted tier, and several projects have no paid tier at all. Two entries break that pattern in opposite directions.\nmemento-mcp\nis listed as free and it is, in the sense that the code costs nothing. Running it requires your own Neo4j 5.13+ instance and a paid OpenAI key for embeddings. The bill exists. It is just not on the invoice you were reading.\nMemori is the outlier the other way. Its first paid step is quoted from\n$60,000 a year\n, with nothing self-serve between the free open-source path and that. It is not a more expensive version of the same purchase. It is a different purchase.\nThe distinction worth holding: free and self-hosted are not the same property.\nmcp-memory-service\nis both, with local embeddings and no cloud bill.\nmemento-mcp\nis neither, once you count what it needs to run.\nThe one filter that actually narrows the field\nAfter all thirteen, the question that did the most work was not on any feature grid.\nDoes memory need to stay on one machine you control, or must several clients and users share one store?\nThat single question splits the field before any other comparison applies, because it sets the trust boundary and the operating burden. Local-first means backups, versioning, and portability become your job. Hosted means the agent's memory lives on someone else's infrastructure.\nAnd one caution inside the local branch that took me too long to see.\nbasic-memory\nconnects to many client applications, Obsidian included. That is a genuine strength, and it is often read as multi-user support. It is not. It serves one user's files across many clients.\n\"Works with many clients\" and \"isolates many users\" are different properties\n, and the second one is a deployment concern in every hosted case here, never an automatic guarantee.\nWhat I would check now instead\nFour questions, all faster than reading a feature comparison:\nWhich repository does the star count belong to?\nOpen the one the install command points at.\nWhen was the last commit to main?\nNot the last release. The last commit.\nDoes the sub-project in the README still exist?\nSearch the repository for it.\nWhat does \"free\" require you to run?\nA database and an embeddings key are a bill.\nNone of that tells you which server is best. Nothing does, because the answer depends on the deployment question above. What it does is stop you shortlisting on numbers that describe something else.\nThe full comparison of all thirteen, with three tables and every number dated and sourced, is on the\noriginal page\n. One of the thirteen is ours. It is flagged as a vendor entry there and ordered first on purpose, so it can be discounted in one glance rather than quietly benefiting from the company it keeps.\nDisclosure: I work on\nMnemoverse\n, a memory engine for AI agents connected over MCP, and it is one of the thirteen servers in the comparison this article draws on. Every number above is dated and checkable on the source page. Free key:\nconsole.mnemoverse.com\n·\nPlans and limits", "metadata": { "title": "I Opened All Thirteen Memory MCP Servers. Every Public Signal I Trusted Was Wrong. - DEV Community", "description": "I went looking for a memory MCP server the way anyone does. Sort by stars, check the registry, read... Tagged with ai, mcp, agents, opensource.", "keywords": "ai, mcp, agents, opensource, software, coding, development, engineering, inclusive, community", "author": null, "og:type": "article", "og:url": "https://dev.to/izgorodin/i-opened-all-thirteen-memory-mcp-servers-every-public-signal-i-trusted-was-wrong-1i1g", "og:title": "I Opened All Thirteen Memory MCP Servers. Every Public Signal I Trusted Was Wrong.", "og:description": "I went looking for a memory MCP server the way anyone does. Sort by stars, check the registry, read...", "og:site_name": "DEV Community", "og:image": "https://media2.dev.to/dynamic/image/width=1200,height=627,fit=cover,gravity=auto,format=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fthwbb69vop1mgzrfwdvq.png", "twitter:site": "@thepracticaldev", "twitter:creator": "@", "twitter:title": "I Opened All Thirteen Memory MCP Servers. Every Public Signal I Trusted Was Wrong.", "twitter:description": "I went looking for a memory MCP server the way anyone does. Sort by stars, check the registry, read...", "twitter:card": "summary_large_image", "twitter:widgets:new-embed-design": "on", "twitter:image:src": "https://media2.dev.to/dynamic/image/width=1200,height=627,fit=cover,gravity=auto,format=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fthwbb69vop1mgzrfwdvq.png" }, "paywall": { "listed_paid": false, "detected": false, "marker": "", "public_only": false, "policy": "付费/会员内容只保留页面公开可见部分，不绕过付费墙" }, "crawl": { "success": true, "status_code": 200, "error": "" }, "code_blocks": [], "images": [], "audio": [], "videos": [], "links": [ { "href": "https://mnemoverse.com/docs/library/memory-mcp-servers-compared", "label": "original page", "type": "a" }, { "href": "https://mnemoverse.com/", "label": "Mnemoverse", "type": "a" }, { "href": "https://console.mnemoverse.com/", "label": "console.mnemoverse.com", "type": "a" }, { "href": "https://mnemoverse.com/pricing", "label": "Plans and limits", "type": "a" } ], "repair_version": 2, "repaired_at": "2026-10-14T10:49:31Z", "source_url": "https://dev.to/izgorodin/i-opened-all-thirteen-memory-mcp-servers-every-public-signal-i-trusted-was-wrong-1i1g", "papers": [], "actual_video_files": [], "body_chars": 7965, "expected_body_images": 0, "discovery_status": "ok", "delivery_status": "ok", "missing_reason": "", "raw_html": "attachments/6c7a7a238625/pages/source.html", "local_html": "attachments/6c7a7a238625/pages/article.html", "body_base_path": "attachments/6c7a7a238625/pages/", "body_markdown_path": "attachments/6c7a7a238625/pages/article.md" }