Skip to main content
MCP server CX Codex

SynaBun — Claude Code MCP server: persistent memory, how to install

SynaBun stores what the agent decided, fixed and explained in a local SQLite file. The next session picks up where the last one left off.

Apache-2.0 ★ 75 updated 2026-10-29

Install into Claude Code · Platform CLI All methods and platforms ↓
claude mcp add synabun -- npx -y synabun
Open on GitHub

What SynaBun does

SynaBun gives Claude Code, Codex and other MCP clients memory tools: remember, recall, reflect, forget. The agent writes facts during work and searches them before the next task — by keywords and by meaning, through an embeddings model that runs locally. Everything lives in one memory.db file; no accounts, cloud services or API keys.

It ships with the Neural Interface, a web app at localhost:3344 that shows the memory map, terminals, a file explorer and side panels for running agents. The server registers itself in Claude Code and Codex with one command; for Cursor, Windsurf and other MCP clients you paste a config snippet. Needs Node.js 22.19 or later. After global install, synabun starts the interface, synabun mcp runs the stdio server for MCP clients.

What it helps with

  • keep context across agent sessions in different projects
  • search past decisions by meaning, not just keywords
  • see and edit what the agent remembered
  • run Claude Code, Codex and OpenCode from one interface with shared memory

Example prompts for the agent

  • Remember we deploy from the release branch, never from main
  • What do you recall about how we deploy?
  • Find all facts about database settings

How to install SynaBun

Platform CLI

claude mcp add synabun -- npx -y synabun

Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more

Questions about SynaBun

How to install SynaBun in Claude Code?

Install globally with npm install -g synabun, then register in Claude Code by running claude mcp add synabun -- npx -y synabun. The agent will get access to memory tools.

What does SynaBun do?

It saves decisions, fixes and explanations from agent sessions to a local SQLite database and retrieves them on request in later sessions. Search works by keywords and by meaning through an embeddings model that runs on your machine.

Why run setup:embeddings?

It downloads the model for search by meaning. Without it SynaBun only searches by keywords.

Similar MCP servers

All MCP servers for Agent setup: memory, context, orchestration →

Project sponsors

Companies that keep this analytics open