Skip to content
Back to search

supermemory

Memory and context engine + app that is extremely fast, scalable, and can be run fully locally. The Memory API for the AI era.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

It has changed since its scan on 11 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/supermemoryai/supermemory with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/supermemoryai/supermemory'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/supermemoryai--supermemory/badge.svg)](/listings/supermemoryai--supermemory)

HTML

<a href="/listings/supermemoryai--supermemory"><img src="/api/listings/supermemoryai--supermemory/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

npm install supermemory    # or: pip install supermemory

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 432 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. <p align="center">
  2. <picture>
  3. <source srcset="apps/web/public/logo-fullmark.svg" media="(prefers-color-scheme: dark)">
  4. <source srcset="apps/web/public/logo-light-fullmark.svg" media="(prefers-color-scheme: light)">
  5. <img src="apps/web/public/logo-fullmark.svg" alt="Supermemory" width="400" />
  6. </picture>
  7. </p>
  8. <p align="center">
  9. <strong>State-of-the-art memory and context engine for AI.</strong>
  10. </p>
  11. <p align="center">
  12. <a href="https://supermemory.ai/docs">Docs</a> ·
  13. <a href="https://supermemory.ai/docs/quickstart">Quickstart</a> ·
  14. <a href="https://supermemory.ai/docs/self-hosting/overview">Self-host</a> ·
  15. <a href="https://console.supermemory.ai">Dashboard</a> ·
  16. <a href="https://supermemory.link/discord">Discord</a>
  17. </p>
  18. <p align="center">
  19. <a href="https://www.npmjs.com/package/supermemory"><img src="https://img.shields.io/npm/v/supermemory?style=flat-square&color=blue" alt="npm" /></a>
  20. <a href="https://pypi.org/project/supermemory/"><img src="https://img.shields.io/pypi/v/supermemory?style=flat-square&color=blue" alt="pypi" /></a>
  21. <a href="https://supermemory.ai/docs"><img src="https://img.shields.io/badge/docs-supermemory.ai-blue?style=flat-square" alt="docs" /></a>
  22. </p>
  23. <p align="center">
  24. <strong>English</strong> · <a href="README.zh-CN.md">简体中文</a>
  25. </p>
  26. <p align="center">
  27. <strong>#1 on every major AI memory benchmark — <a href="https://github.com/xiaowu0162/LongMemEval">LongMemEval</a>, <a href="https://github.com/snap-research/locomo">LoCoMo</a>, and <a href="https://github.com/Salesforce/ConvoMem">ConvoMem</a>.</strong><br/>
  28. <strong>95% Recall@15 with a 99.4% context reduction · ~50ms user profiles.</strong><br/>
  29. <a href="https://supermemory.ai/research">Read the research →</a>
  30. </p>
  31. ---
  32. Supermemory is the memory and context layer for AI. **#1 on [LongMemEval](https://github.com/xiaowu0162/LongMemEval), [LoCoMo](https://github.com/snap-research/locomo), and [ConvoMem](https://github.com/Salesforce/ConvoMem)** — the three major benchmarks for AI memory.
  33. We are a research lab building the engine, plugins and tools around it.
  34. Your AI forgets everything between conversations. Supermemory fixes that.
  35. It automatically learns from conversations, extracts facts, builds user profiles, handles knowledge updates and contradictions, forgets expired information, and delivers the right context at the right time. Full RAG, connectors, file processing — the entire context stack, one system.
  36. | | |
  37. |---|---|
  38. | 🧠 **Memory** | Extracts facts from conversations. Handles temporal changes, contradictions, and automatic forgetting. |
  39. | 👤 **User Profiles** | Auto-maintained user context — stable facts + recent activity. One call, ~50ms. |
  40. | 🔍 **Hybrid Search** | RAG + Memory in a single query. Knowledge base docs and personalized context together. |
  41. | 🔌 **Connectors** | Google Drive · Gmail · Notion · OneDrive · GitHub — auto-sync with real-time webhooks. |
  42. | 📄 **Multi-modal Extractors** | PDFs, images (OCR), videos (transcription), code (AST-aware chunking). Upload and it works. |
  43. All of this is in our single memory structure and ontology.
  44. <img width="1414" height="937" alt="image" src="https://github.com/user-attachments/assets/8863b6d9-c043-4c75-b200-4f1759e7edaf" />
  45. ---
  46. ## Use Supermemory
  47. <table>
  48. <tr>
  49. <td width="50%" valign="top">
  50. <h3>🧑‍💻 I use AI tools</h3>
  51. Give Claude Code, Muse Code, Cursor, Codex and OpenCode **persistent memory across every conversation** with a plugin or the MCP server.
  52. Your AI remembers your preferences, projects, past discussions — and gets smarter over time.
  53. **[→ Jump to User setup](#give-your-ai-memory)**
  54. </td>
  55. <td width="50%" valign="top">
  56. <h3>🔧 I'm building AI products</h3>
  57. Add memory, RAG, user profiles, and connectors to your agents and apps with **a single API**.
  58. No vector DB config. No embedding pipelines. No chunking strategies.
  59. **[→ Jump to developer quickstart](#build-with-supermemory-api)**
  60. </td>
  61. </tr>
  62. <tr>
  63. <td colspan="2" valign="top">
  64. <h3>🖥️ I want to run it myself</h3>
  65. State-of-the-art memory, on your machine. **One binary. Zero config.** Bring any model — or run fully offline with Ollama.
  66. ```bash
  67. curl -fsSL https://supermemory.ai/install | bash
  68. ```
  69. **[→ Jump to Supermemory local](#supermemory-local--run-it-yourself)**
  70. </td>
  71. </tr>
  72. </table>
  73. ---
  74. ## Give your AI memory
  75. Plugins and the MCP server give any compatible AI assistant persistent memory. One install, and your AI remembers you.
  76. ### Supermemory Plugins
  77. Supermemory comes built with plugins for Claude Code, Muse Code, Cursor, Codex, OpenCode, OpenClaw, and Hermes.
  78. <img width="844" height="484" alt="image" src="https://github.com/user-attachments/assets/ecb879a2-8652-495d-9228-f305a97ba603" />
  79. These plugins are implementations of the supermemory API, and they are open source!
  80. You can find them here:
  81. - Claude Code plugin: https://github.com/supermemoryai/claude-supermemory
  82. - Muse Code plugin: https://github.com/supermemoryai/muse-supermemory
  83. - Cursor plugin: https://github.com/supermemoryai/cursor-supermemory
  84. - Codex plugin: https://github.com/supermemoryai/codex-supermemory
  85. - OpenClaw plugin: https://github.com/supermemoryai/openclaw-supermemory
  86. - OpenCode plugin: https://github.com/supermemoryai/opencode-supermemory
  87. - Hermes agent (Supermemory memory provider): https://github.com/NousResearch/hermes-agent
  88. ### MCP
  89. Server URL:
  90. ```text
  91. https://mcp.supermemory.ai/mcp
  92. ```
  93. ```json
  94. {
  95. "mcpServers": {
  96. "supermemory": {
  97. "url": "https://mcp.supermemory.ai/mcp"
  98. }
  99. }
  100. }
  101. ```
  102. Read more about our MCP here - https://supermemory.ai/docs/supermemory-mcp/mcp
  103. ### What your AI gets
  104. | Tool | What it does |
  105. |---|---|
  106. | `memory` | Save or forget information. Your AI calls this automatically when you share something worth remembering. |
  107. | `recall` | Search memories by query. Returns relevant memories + your user profile summary. |
  108. | `context` | Injects your full profile (preferences, recent activity) into the conversation at start. In Cursor and Claude Code, just type `/context`. |
  109. ### How it works
  110. Once installed, Supermemory runs in the background:
  111. 1. **You talk to your AI normally.** Share preferences, mention projects, discuss problems.
  112. 2. **Supermemory extracts and stores the important stuff.** Facts, preferences, project context — not noise.
  113. 3. **Next conversation, your AI already knows you.** It recalls what you're working on, how you like things, what you discussed before.
  114. Memory is scoped with **projects** (container tags) so you can separate work and personal context, or organize by client, repo, or anything else.
  115. ### Supported clients
  116. **Claude Desktop** · **Cursor** · **Windsurf** · **VS Code** · **Claude Code** · **OpenCode** · **OpenClaw** · **Hermes**
  117. The MCP server is open source — [view the source](https://supermemory.ai/docs/supermemory-mcp/mcp).
  118. ### Manual configuration
  119. Add this to your MCP client config:
  120. ```json
  121. {
  122. "mcpServers": {
  123. "supermemory": {
  124. "url": "https://mcp.supermemory.ai/mcp"
  125. }
  126. }
  127. }
  128. ```
  129. ---
  130. ## Build with Supermemory (API)
  131. If you're building AI agents or apps, Supermemory gives you the entire context stack through one API — memory, RAG, user profiles, connectors, and file processing.
  132. ### Install
  133. ```bash
  134. npm install supermemory # or: pip install supermemory
  135. ```
  136. ### Quickstart
  137. ```typescript
  138. import Supermemory from "supermemory";
  139. const client = new Supermemory();
  140. // Store a conversation
  141. await client.add({
  142. content: "User loves TypeScript and prefers functional patterns",
  143. containerTag: "user_123",
  144. });
  145. // Get user profile + relevant memories in one call
  146. const { profile, searchResults } = await client.profile({
  147. containerTag: "user_123",
  148. q: "What programming style does the user prefer?",
  149. });
  150. // profile.static → ["Loves TypeScript", "Prefers functional patterns"]
  151. // profile.dynamic → ["Working on API integration"]
  152. // searchResults → Relevant memories ranked by similarity
  153. ```
  154. ```python
  155. from supermemory import Supermemory
  156. client = Supermemory()
  157. client.add(
  158. content="User loves TypeScript and prefers functional patterns",
  159. container_tag="user_123"
  160. )
  161. result = client.profile(container_tag="user_123", q="programming style")
  162. print(result.profile.static) # Long-term facts
  163. print(result.profile.dynamic) # Recent context
  164. ```
  165. Supermemory automatically extracts memories, builds user profiles, and returns relevant context. No embedding pipelines, no vector DB config, no chunking strategies.
  166. ### Framework integrations
  167. Drop-in wrappers for every major AI framework:
  168. ```typescript
  169. // Vercel AI SDK
  170. import { withSupermemory } from "@supermemory/tools/ai-sdk";
  171. const model = withSupermemory(openai("gpt-4o"), { containerTag: "user_123", customId: "conv-1" });
  172. // Mastra
  173. import { withSupermemory } from "@supermemory/tools/mastra";
  174. const agent = new Agent(withSupermemory(config, "user-123", { mode: "full" }));
  175. ```
  176. **Vercel AI SDK** · **LangChain** · **LangGraph** · **OpenAI Agents SDK** · **Mastra** · **Agno** · **Claude Memory Tool** · **n8n**
  177. ### Search modes
  178. ```typescript
  179. // Hybrid (default) — RAG + Memory in one query
  180. const results = await client.search({
  181. q: "how do I deploy?",
  182. containerTag: "user_123",
  183. searchMode: "hybrid",
  184. });
  185. // Returns deployment docs (RAG) + user's deploy preferences (Memory)
  186. // Memories only
  187. const results = await client.search({
  188. q: "user preferences",
  189. containerTag: "user_123",
  190. searchMode: "memories",
  191. });
  192. ```
  193. ### User profiles
  194. Traditional memory relies on search — you need to know what to ask for. Supermemory automatically maintains a profile for every user:
  195. ```typescript
  196. const { profile } = await client.profile({ containerTag: "user_123" });
  197. // profile.static → ["Senior engineer at Acme", "Prefers dark mode", "Uses Vim"]
  198. // profile.dynamic → ["Working on auth migration", "Debugging rate limits"]
  199. ```
  200. One call. ~50ms. Inject into your system prompt and your agent instantly knows who it's talking to.
  201. ### Connectors
  202. Auto-sync external data into your knowledge base:
  203. **Google Drive** · **Gmail** · **Notion** · **OneDrive** · **GitHub** · **Web Crawler**
  204. Real-time webhooks. Documents automatically processed, chunked, and searchable.
  205. ### API at a glance
  206. | Method | Purpose |
  207. |---|---|
  208. | `client.add()` | Store content — text, conversations, URLs, HTML |
  209. | `client.profile()` | User profile + optional search in one call |
  210. | `client.search()` | Hybrid search across memories and documents (`searchMode`) |
  211. | `client.search.documents()` | Document search with metadata filters (legacy v3 response shape) |
  212. | `client.documents.uploadFile()` | Upload PDFs, images, videos, code |
  213. | `client.documents.list()` | List and filter documents |
  214. | `client.settings.update()` | Configure memory extraction and chunking |
  215. Full API reference → [supermemory.ai/docs](https://supermemory.ai/docs)
  216. ---
  217. ## Supermemory local — run it yourself
  218. State-of-the-art memory, on your machine. One binary. Zero config.
  219. ```bash
  220. curl -fsSL https://supermemory.ai/install | bash
  221. # or
  222. npx supermemory local
  223. ```
  224. ```bash
  225. supermemory-server
  226. ```
  227. First boot sets up the embedded Supermemory graph engine, local embeddings, and your credentials, then prints an API key. The full Memory API — documents, memories, user profiles, hybrid search — runs against `http://localhost:6767`.
  228. ```typescript
  229. const client = new Supermemory({
  230. apiKey: "sm_...",
  231. baseURL: "http://localhost:6767", // that's the only change
  232. });
  233. ```
  234. - **Bring any model** — OpenAI, Anthropic, Gemini, Groq, or any OpenAI-compatible endpoint. An interactive wizard walks you through it on first boot.
  235. - **Embeddings** — local `Xenova/bge-base-en-v1.5` by default (no API key); optionally OpenAI, Gemini, or Ollama. Same provider stack as cloud.
  236. - **Offline text memories if you want** — point it at Ollama (`gpt-oss:20b` works great), disable telemetry with `SUPERMEMORY_DISABLE_TELEMETRY=1`, and use local text or file inputs after the first embedding-model download. URL ingestion uses a hosted reader service.
  237. - **Your data, one directory** — everything lives in `~/.supermemory`, easy to back up or move.
  238. - **Same API as the platform** — prototype locally, ship on the hosted platform by changing `baseURL`.
  239. ### Run with Docker
  240. Prebuilt images support Linux amd64 and arm64. You don’t need to build or clone the server source.
  241. For a new installation, run these commands in an empty directory and set your own LLM provider key:
  242. ```bash
  243. release_url=https://github.com/supermemoryai/supermemory/releases
  244. curl -fL "$release_url/download/server-v0.0.9/compose.yaml" -o compose.yaml
  245. export OPENAI_API_KEY=your_openai_api_key_here
  246. export SUPERMEMORY_VERSION=0.0.9
  247. docker compose up -d --wait --wait-timeout 180
  248. ```
  249. The API listens on `http://localhost:6767`. Data and credentials persist in the `supermemory_supermemory-data` volume. Get your local API key with `docker compose exec supermemory cat /data/api-key`; requests require it even on localhost.
  250. Read the [Docker setup and upgrade instructions](apps/docs/self-hosting/docker.mdx) for persistent provider configuration, backups and upgrades. Keep the existing volume and keys when upgrading; **don’t run `docker compose down -v`**.
  251. Read the [self-hosting docs](https://supermemory.ai/docs/self-hosting/overview) — [quickstart](https://supermemory.ai/docs/self-hosting/quickstart), [configuration](https://supermemory.ai/docs/self-hosting/configuration), [embeddings](https://supermemory.ai/docs/self-hosting/embeddings), and [local vs. Enterprise](https://supermemory.ai/docs/self-hosting/local-vs-enterprise).
  252. ---
  253. ## Benchmarks
  254. Supermemory is state of the art across all major AI memory benchmarks:
  255. | Benchmark | What it measures | Result |
  256. |---|---|---|
  257. | **[LongMemEval](https://github.com/xiaowu0162/LongMemEval)** | Long-term memory across sessions with knowledge updates | **#1** |
  258. | **[LoCoMo](https://github.com/snap-research/locomo)** | Fact recall across extended conversations (single-hop, multi-hop, temporal, adversarial) | **#1** |
  259. | **[ConvoMem](https://github.com/Salesforce/ConvoMem)** | Personalization and preference learning | **#1** |
  260. On LongMemEval, supermemory reaches **95% Recall@15 while adding only ~720 tokens of context — a 99.4% context reduction** (99.6% at @10, 99.8% at @5). Recall by category: Knowledge Updates 99%, Assistant recall 100%, User recall 97%, Multi-session 93%, Temporal Reasoning 91%, Preference 90%.
  261. We also built the **Supermemory Filesystem (SMFS)**, which uses **3.0× fewer tokens on Claude** (24M vs 72M) and **1.75× fewer on Codex** across the 110-question xAFS benchmark. See the full write-ups on our [research page](https://supermemory.ai/research).
  262. We also built **[MemoryBench](https://supermemory.ai/docs/memorybench/overview)** — an open-source framework for standardized, reproducible benchmarks of memory providers. Compare Supermemory, Mem0, Zep, and others head-to-head:
  263. ```bash
  264. bun run src/index.ts run -p supermemory -b longmemeval -j gpt-4o -r my-run
  265. ```
  266. ### Benchmarking your own memory solution
  267. We provide an Agent skill for companies to benchmark their own context and memory solutions against supermemory.
  268. ```
  269. npx skills add supermemoryai/memorybench
  270. ```
  271. Simply run this and do `/benchmark-context` - Supermemory will automatically do the work for you!
  272. ---
  273. ## How memory works under the hood
  274. ```
  275. Your app / AI tool
  276. ↓
  277. Supermemory
  278. │
  279. ├── Memory Engine Extracts facts, tracks updates, resolves contradictions,
  280. │ auto-forgets expired info
  281. ├── User Profiles Static facts + dynamic context built from engine, always fresh
  282. ├── Hybrid Search RAG + Memory in one query
  283. ├── Connectors Real-time sync from Google Drive, Gmail, Notion, GitHub...
  284. └── File Processing PDFs, images, videos, code → searchable chunks
  285. ```
  286. **Memory is not RAG.** RAG retrieves document chunks — stateless, same results for everyone. Memory extracts and tracks *facts about users* over time. It understands that "I just moved to SF" supersedes "I live in NYC." Supermemory runs both together by default, so you get knowledge base retrieval *and* personalized context in every query. Read more about this here - https://supermemory.ai/docs/concepts/memory-vs-rag
  287. **Automatic forgetting.** Supermemory knows when memories become irrelevant. Temporary facts ("I have an exam tomorrow") expire after the date passes. Contradictions are resolved automatically. Noise never becomes permanent memory.
  288. ---
  289. ## Links
  290. - 📖 [Documentation](https://supermemory.ai/docs)
  291. - 🚀 [Quickstart](https://supermemory.ai/docs/quickstart)
  292. - 🖥️ [Self-hosting (Supermemory local)](https://supermemory.ai/docs/self-hosting/overview)
  293. - 🧪 [MemoryBench](https://supermemory.ai/docs/memorybench/overview)
  294. - 🔌 [Integrations](https://supermemory.ai/docs/integrations)
  295. - 💬 [Discord](https://supermemory.link/discord)
  296. - 𝕏 [Twitter](https://twitter.com/supermemory)
  297. ---
  298. <p align="center">
  299. <strong>Give your AI a memory. It's about time..</strong>
  300. </p>

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.