Persistent, project-scoped memory for coding agents, with local storage, recall, contradiction detection, and audit history.
Copy the AI prompt to install this server into Claude Code, Cursor, or another agent — or use 1-click editor setup below.
The install command below didn't complete successfully in our automated test.
npx -y timps-codenpm notice npm notice New major version of npm available! 10.1.0 -> 12.1.0 npm notice Changelog: <https://github.com/npm/cli/releases/tag/v12.1.0> npm notice Run `npm install -g npm@12.1.0` to update! npm notice npm ERR! code E404 npm ERR! 404 Not Found - GET https://registry.npmjs.org/@timps%2fmemory-core - Not found npm ERR! 404 npm ERR! 404 '@timps/memory-core@*' is not in this registry. npm ERR! 404 npm ERR! 404 Note that you can also install from a npm ERR! 404 tarball, folder, http ur
This is an experimental automated check and can have false negatives — missing environment variables, a slow cold install, etc. It doesn’t necessarily mean something’s wrong. Last checked 13d ago.
💡 Paste the JSON block into your client's configuration file under mcpServers, then restart the application.
Inspect callable tools, capabilities, and parameters exposed to AI agents by Timps.
The timps MCP server adds persistent memory to coding agents without replacing the agent itself. It keeps a separate memory store for each project, so information from different repositories does not normally mix. The stored information can include architectural decisions, codebase-specific conventions, previously encountered bugs and their fixes, session outcomes, and an audit record explaining what was written.
Memory is divided into several layers. Working memory covers the current session and is cleared when the session ends. Episodic memory records previous sessions and their outcomes. Semantic memory holds durable facts such as patterns, conventions, and decisions promoted from past activity. An append-only, hash-chained audit trail records memory writes and when they occurred.
Additional processing includes temporal pattern detection, contradiction detection, confidence scoring, and importance-based pruning. These features help the memory store identify conflicting information and retain more relevant durable context.
The MCP server runs between an MCP-compatible coding agent and the TIMPS memory store. A client starts the server over stdio, then the agent can use its memory capabilities during work. On later sessions, the project-specific store can provide context from earlier work instead of requiring the user to repeat it.
Local JSON and SQLite storage are the default options. Postgres and Qdrant are available as optional storage choices when a shared or externally hosted setup is needed. Ollama can provide local model processing, while Claude, GPT, and Gemini are listed as optional provider choices. The README does not specify the configuration details for switching between these providers.
Install the MCP package globally with:
Then configure the client to launch timps-mcp as a stdio process. The README provides configurations for Claude Code, Cursor, and Windsurf, each using the same command. Other MCP clients can also run the command over stdio.
No API key is required when using the local Ollama path. The project also supports optional paid model providers, so provider-specific credentials may be needed when those services are selected. The supplied material does not list environment variable names or detailed storage configuration options.
The timps MCP server supports these documented capabilities:
The repository describes the MCP server and CLI as the parts used and trusted most by the author, while the desktop app, mobile app, JetBrains plugin, VS Code extension, and swarm mode are identified as less tested. Treat those additional components separately from the core MCP use case.
Published benchmark results are self-run on a synthetic corpus and are not independently verified. They should not be treated as performance guarantees for a particular codebase. The README also does not provide a complete tool list, detailed provider setup, or environment-variable reference. The timps MCP server is therefore best evaluated against the memory and client workflow your project requires.
Factual signals from GitHub, npm, and our automated checks — not a rating.
No reviews yet — be the first to share how this listing worked for you.
Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.
[](https://allmcps.com/mcp/timps)<a href="https://allmcps.com/mcp/timps"><img src="https://allmcps.com/api/badge/timps?style=directory" alt="Timps on AllMCPs" /></a>