Memory Custom : PouchDB

Memory Custom : PouchDB is an MCP server that integrates PouchDB document storage into the Model Context Protocol memory server. Built as a fork of the custom memory server, it connects language models like Claude to a persistent local database rather than relying exclusively on a growing JSON file. Developers, researchers, and AI engineers use this tool to manage dynamic knowledge graphs across long-running sessions. The server enables users to store structured entities, relations, and observations with interaction timestamps. By leveraging PouchDB, it improves data consistency, introduces document versioning, and provides better performance for larger datasets. It also supports custom database and backup file paths, making it possible to isolate project-specific memories. Users can optionally disable the default JSON backup to rely entirely on PouchDB. With dedicated tools for graph reading, node searching, and relation tracking, the server allows AI assistants to reliably recall past interactions, update user profiles, and maintain coherent multi-turn context over time.

Category: AI Memory & Context

Tags: memory, persistence, pouchdb, storage

Visit Memory Custom : PouchDB

How to install and configure Memory Custom : PouchDB

1. Ensure Node.js (version 16 or higher) is installed. 2. Clone the repository and install dependencies: bash git clone git@github.com:bneil/mcp-memory-pouchdb.git cd mcp-memory-pouchdb npm install npm run build Alternatively, install via Smithery: bash npx -y @smithery/cli install @bneil/mcp-memory-pouchdb --client claude 3. Configure the server in your claude_desktop_config.json or .cursor/mcp.json file: json { "mcpServers": { "memory": { "command": "node", "args": ["/path/to/mcp-memory-pouchdb/dist/index.js"], "env": { "MEMORY_FILE_PATH": "/path/to/custom/memory.json", "POUCHDB_PATH": "/path/to/custom/pouchdb_directory", "DISABLE_MEMORY_FILE": "true" } } } } 4. Restart your MCP client to initialize the server.

What you can do with Memory Custom : PouchDB

• Storing and recalling user preferences, basic identity, and interaction history across Claude conversations using persistent PouchDB records. - Managing isolated knowledge graphs for distinct projects by defining unique directory paths via the POUCHDB_PATH variable. - Tracking when observations and entities were created or modified using the integrated interaction timestamping feature. - Querying, opening, and navigating complex relationship networks between people, organizations, and concepts during research. - Preventing performance degradation caused by massive single-file JSON memories through efficient document-based storage indexing.

Key facts

• Open Source
• https://github.com/bneil/mcp-memory-pouchdb
• AI Memory & Context, Databases & Data Stores
• memory, persistence, pouchdb, storage

Part of MCP Servers

How do I install Memory Custom : PouchDB?

You can install it automatically via the Smithery CLI using npx -y @smithery/cli install @bneil/mcp-memory-pouchdb --client claude. Alternatively, clone the repository from GitHub, run npm install followed by npm run build, and register the built index.js script within your MCP client configuration file.

What can Memory Custom : PouchDB do?

It manages a persistent knowledge graph using PouchDB for document storage. The server provides tools to create, read, and delete entities, relations, and observations. It also adds interaction timestamps, searches nodes, and allows custom storage locations for both database files and JSON backups.

Which MCP clients work with Memory Custom : PouchDB?

The server works with any Model Context Protocol client that supports standard input/output server processes and environment variables, including Claude Desktop and Cursor. Users configure the server by adding the startup command and required paths to their client configuration file.

Which environment variables are required?

The server requires two main environment variables: MEMORY_FILE_PATH, which defines the path for the backup JSON file, and POUCHDB_PATH, which sets the database storage directory. Optional variables include POUCHDB_OPTIONS for configuration settings and DISABLE_MEMORY_FILE to disable JSON backups.

Is Memory Custom : PouchDB open source?

Yes, Memory Custom : PouchDB is open source. The repository is publicly accessible on GitHub at https://github.com/bneil/mcp-memory-pouchdb, allowing users to review the codebase, inspect updates, or fork the repository for custom enhancements.