Persistent memory for Claude Code, Cursor and Codex. Facts are typed, sourced, and retired when they change. Local SQLite, no API keys.
- ✓Open-source license (Apache-2.0)
- ✓Actively maintained (<30d)
- ✓Clear description
- ✓Topics declared
- ✓Documented (README)
claude mcp add knowl -- npx -y @dat999zx/knowl{
"mcpServers": {
"knowl": {
"command": "npx",
"args": ["-y", "@dat999zx/knowl"]
}
}
}Resumen de MCP Servers
<div align="center"> <img src="docs/assets/hero.png" alt="Knowl — the knowledge operating system for AI agents: memory, knowledge, context, continuity" width="100%" /> **Your CLAUDE.md only grows. Knowl retires facts when they change.** [](https://www.npmjs.com/package/@dat999zx/knowl) [](https://github.com/dat999zx/knowl/actions/workflows/ci.yml) [](LICENSE) [](package.json) [](https://modelcontextprotocol.io) <a href="https://discord.gg/YrWbHe8FBB"><img src="https://img.shields.io/badge/Discord-Join%20the%20server-5865F2?logo=discord&logoColor=white&style=for-the-badge" alt="Join the Knowl Discord" height="40" /></a> <p align="center"> <a href="#the-idea-memory-that-retires-itself"><picture><source media="(prefers-color-scheme: light)" srcset="docs/assets/chips/light/stat-supersession.svg"><img src="docs/assets/chips/stat-supersession.svg" alt="Scores 90 on MemoryAgentBench FactConsolidation single-hop at 262K" height="38" /></picture></a> <a href="#quick-start"><picture><source media="(prefers-color-scheme: light)" srcset="docs/assets/chips/light/stat-nokeys.svg"><img src="docs/assets/chips/stat-nokeys.svg" alt="0 API keys needed" height="38" /></picture></a> <a href="#everything-else"><picture><source media="(prefers-color-scheme: light)" srcset="docs/assets/chips/light/stat-tools.svg"><img src="docs/assets/chips/stat-tools.svg" alt="28 MCP tools" height="38" /></picture></a> <a href="#what-knowl-is-for"><picture><source media="(prefers-color-scheme: light)" srcset="docs/assets/chips/light/stat-local.svg"><img src="docs/assets/chips/stat-local.svg" alt="100% local, no egress" height="38" /></picture></a> </p> [Quick start](#quick-start) · [Why supersession](#the-idea-memory-that-retires-itself) · [What gets stored](#what-gets-stored) · [Features](#features) · [Agent setup](#connecting-an-agent) · [Viewer](#see-it-the-local-viewer) · [Requirements](#requirements-and-local-data) · **[Full reference →](docs/reference.md)** </div> --- Your agent starts every session blank, so you keep a `CLAUDE.md`. It only grows. Six months in it still names the database you migrated off last spring, and now the agent gets both answers. Knowl is persistent memory for Claude Code, Cursor and Codex, over [MCP](https://modelcontextprotocol.io) or the CLI. When a fact is replaced, the old one is retired instead of competing with the new one. No API key needed. When Knowl isn't sure the new fact replaces the old, it leaves both active and hands you the `knowl supersede` command to say so. Turn that off and retrieval drops from 98% to 47%. End to end, 90 to 73. [How it was measured ↓](#the-idea-memory-that-retires-itself) Forty seconds, one decision, three agents: <div align="center"> <img src="docs/assets/knowl-showcase.webp" alt="Claude Code answers which database the project uses from memory, records the move to Postgres and retires the MySQL decision; Codex answers the same question from that memory in a second terminal; the Claude app answers from the same store over the hosted connector" width="88%" /> </div> ## Quick start Requires Node.js 22 or later. macOS, Linux and Windows. ```bash npm install -g @dat999zx/knowl cd your-project knowl init ``` <details> <summary>Other package managers</summary> The published package is the same one in every case; each of these installs it and puts `knowl` on your `PATH`. ```bash pnpm add -g @dat999zx/knowl yarn global add @dat999zx/knowl bun add -g @dat999zx/knowl ``` Or run it without installing: ```bash npx @dat999zx/knowl init ``` Knowl runs on Node.js in all of these — Bun installs it, Node executes it. It bundles native addons (SQLite, tree-sitter, the embedding runtime), so running the CLI under the Bun or Deno runtime directly is not supported. </details> `knowl init` creates `.knowl/`, installs the project guidance files, updates `.gitignore`, and registers Knowl with whichever agents it detects. It also warms a local embedding model (~53 MB) in the background — `init` succeeds either way, and without it you still get keyword search. That is the whole setup. You do not record memory by hand: your agent reads and writes it as it works. ## Connecting an agent <table> <tr> <td align="center" width="14%"> <a href="https://claude.com/product/claude-code"><img src="docs/assets/logos/anthropic.svg" alt="Claude Code" width="40" height="40" /></a><br/> <strong>Claude Code</strong><br/> <sub>MCP · lifecycle · gate</sub> </td> <td align="center" width="14%"> <a href="https://github.com/openai/codex"><img src="docs/assets/logos/openai.svg" alt="Codex" width="40" height="40" /></a><br/> <strong>Codex</strong><br/> <sub>MCP · lifecycle · gate</sub> </td> <td align="center" width="14%"> <a href="https://hermes-agent.nousresearch.com"><img src="docs/assets/logos/hermes.png" alt="Hermes Agent" width="40" height="40" /></a><br/> <strong>Hermes</strong><br/> <sub>MCP · lifecycle · gate</sub> </td> <td align="center" width="14%"> <a href="https://github.com/openclaw/openclaw"><img src="docs/assets/logos/openclaw.svg" alt="OpenClaw" width="40" height="40" /></a><br/> <strong>OpenClaw</strong><br/> <sub>in-process · plugin · gate</sub> </td> <td align="center" width="14%"> <a href="https://github.com/features/copilot"><picture><source media="(prefers-color-scheme: dark)" srcset="docs/assets/logos/githubcopilot-dark.svg" /><img src="docs/assets/logos/githubcopilot.svg" alt="Copilot" width="40" height="40" /></picture></a><br/> <strong>Copilot</strong><br/> <sub>MCP · lifecycle · gate</sub> </td> <td align="center" width="14%"> <a href="https://cursor.com"><picture><source media="(prefers-color-scheme: dark)" srcset="docs/assets/logos/cursor-dark.svg" /><img src="docs/assets/logos/cursor.svg" alt="Cursor" width="40" height="40" /></picture></a><br/> <strong>Cursor</strong><br/> <sub>MCP · lifecycle · gate</sub> </td> <td align="center" width="14%"> <a href="https://openhands.dev"><img src="docs/assets/logos/openhands.svg" alt="OpenHands" width="40" height="40" /></a><br/> <strong>OpenHands</strong><br/> <sub>MCP · lifecycle · gate</sub> </td> </tr> <tr> <td align="center" width="14%"> <a href="https://antigravity.google"><img src="docs/assets/logos/antigravity.svg" alt="Antigravity" width="40" height="40" /></a><br/> <strong>Antigravity</strong><br/> <sub>MCP · lifecycle · gate</sub> </td> <td align="center" width="14%"> <a href="https://windsurf.com"><picture><source media="(prefers-color-scheme: dark)" srcset="docs/assets/logos/windsurf-dark.svg" /><img src="docs/assets/logos/windsurf.svg" alt="Windsurf" width="40" height="40" /></picture></a><br/> <strong>Windsurf</strong><br/> <sub>MCP · lifecycle · gate</sub> </td> <td align="center" width="14%"> <a href="https://github.com/cline/cline"><picture><source media="(prefers-color-scheme: dark)" srcset="docs/assets/logos/cline-dark.svg" /><img src="docs/assets/logos/cline.svg" alt="Cline" width="40" height="40" /></picture></a><br/> <strong>Cline</strong><br/> <sub>MCP · lifecycle · plugin</sub> </td> <td align="center" width="14%"> <a href="https://zed.dev"><img src="docs/assets/logos/zed.svg" alt="Zed" width="40" height="40" /></a><br/> <strong>Zed</strong><br/> <sub>MCP · capture · ACP</sub> </td> <td align="center" width="14%"> <a href="https://www.jetbrains.com"><picture><source media="(prefers-color-scheme: dark)" srcset="docs/assets/logos/jetbrains-dark.svg" /><img src="docs/assets/logos/jetbrains.svg" alt="JetBrains" width="40" height="40" /></picture></a><br/> <strong>JetBrains</strong><br/> <sub>MCP · capture · ACP</sub> </td> <td align="center" width="14%"> <a href="https://github.com/anomalyco/opencode"><picture><source media="(prefers-color-scheme: dark)" srcset="docs/assets/logos/opencode-dark.svg" /><img src="docs/assets/logos/opencode.svg" alt="OpenCode" width="40" height="40" /></picture></a><br/> <strong>OpenCode</strong><br/> <sub>MCP · manual loop</sub> </td> <td align="center" width="14%"> <a href="https://claude.ai/download"><img src="docs/assets/logos/anthropic.svg" alt="Claude Desktop" width="40" height="40" /></a><br/> <strong>Claude Desktop</strong><br/> <sub>MCP · manual loop</sub> </td> </tr> </table> `knowl init` registers the MCP server for every host it finds. Start a new session afterwards so the agent picks up its guidance, and it will query and write memory on its own. **gate** means Knowl can refuse an edit that invalidates code another session is holding. Neovim and Kiro work the same way as Zed and JetBrains, through `knowl acp`. Cline needs one line pointing it at the shipped plugin. Hermes Agent gets a Python plugin, installed for you, that works in the terminal and in Hermes Desktop alike, and can additionally be picked as Hermes' memory provider. OpenClaw runs in-process inside its gateway via an extension plugin, evaluating write gates without subprocess overhead — `knowl init openclaw` copies it and prints the two commands that register it. Any other MCP client works with no integration at all. Running agents in parallel? Every **git worktree** resolves to the main checkout's store — [Conductor](https://conductor.build) workspaces, Claude Code's `isolation: "worktree"`, or your own scripts all share one memory, with nothing to configure. [How that works, and its one limit →](docs/hosts.md#parallel-agents-and-git-worktrees) → [Every host, and what each one can do](docs/hosts.md) · [How agents use it](#how-agents-use-it) · [MCP tools and resources](docs/reference.md#mcp-tools-and-resources) ## The idea: memory that retires itself Most memory systems are append-only. Storing "we moved to SQLite" leaves "we use Postg
Lo que la gente pregunta sobre knowl
¿Qué es dat999zx/knowl?
+
dat999zx/knowl es mcp servers para el ecosistema de Claude AI. Persistent memory for Claude Code, Cursor and Codex. Facts are typed, sourced, and retired when they change. Local SQLite, no API keys. Tiene 84 estrellas en GitHub y su última actualización registrada es del 2026-09-18.
¿Cómo se instala knowl?
+
Puedes instalar knowl clonando el repositorio (https://github.com/dat999zx/knowl) o siguiendo las instrucciones del README en GitHub. ClaudeWave también te ofrece bloques de instalación rápida en esta misma página.
¿Es seguro usar dat999zx/knowl?
+
Nuestro agente de seguridad ha analizado dat999zx/knowl y le ha asignado un Trust Score de 95/100 (tier: Verified). Revisa el desglose completo de comprobaciones superadas y flags en esta página.
¿Quién mantiene dat999zx/knowl?
+
dat999zx/knowl es mantenido por dat999zx. La última actividad registrada en GitHub es del 2026-09-18, con 3 issues abiertos.
¿Hay alternativas a knowl?
+
Sí. En ClaudeWave puedes explorar mcp servers similares en /categories/mcp, ordenados por popularidad o actividad reciente.
Despliega knowl en tu cloud
Lleva este repo a producción en minutos. Cada plataforma genera su propio entorno con variables de entorno editables.
¿Mantienes este repo? Añade un badge a tu README
Pega el badge en tu README de GitHub para mostrar que está auditado por ClaudeWave. Cada badge enlaza de vuelta a esta página y muestra el Trust Score actual.
[](https://claudewave.com/repo/dat999zx-knowl)<a href="https://claudewave.com/repo/dat999zx-knowl"><img src="https://claudewave.com/api/badge/dat999zx-knowl" alt="Featured on ClaudeWave: dat999zx/knowl" width="320" height="64" /></a>Más MCP Servers
Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
An open-source AI agent that brings the power of Gemini directly into your terminal.
Real-time global intelligence dashboard. AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface
🕷️ An adaptive Web Scraping framework that handles everything from a single request to a full-scale crawl! Don't be shy, join here: https://discord.gg/EMgGbDceNQ
The fastest path to AI-powered full stack observability, even for lean teams.