
Rememb
@LuizEduPP
About Rememb
Persistent memory for AI agents — local, portable, zero config. Works with Cursor, Windsurf, Claude via MCP.
Basic information
Category
Developer Tools
License
MIT
Runtime
python
Transports
stdio
Publisher
LuizEduPP
Submitted by
Luiz Eduardo
Config
Add this server to your MCP-compatible client using the configuration below.
{
"mcpServers": {
"rememb": {
"command": "rememb",
"args": [
"mcp"
]
}
}
}Tools
No tools detected
We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.
Overview
What is Rememb?
Rememb is a local-first persistent memory layer for AI agents. It provides structured entries, keyword search, versioning, diff, restore, and an audit trail — all without requiring a cloud service or API keys. It is designed for teams and solo developers who operate agents across sessions and need to avoid context loss.
How to use Rememb?
Install with pip install rememb. The recommended integration is via MCP: add a rememb server entry to your IDE's MCP config using the command rememb mcp. Restart the IDE and the agent can read, write, search, and restore memory automatically. For a persistent local process, start with rememb mcp --transport sse --host 127.0.0.1 --port 8765. You can also run the web UI with rememb to inspect and configure memory.
Key features of Rememb
- Local-first memory with JSON or SQLite storage on disk
- Structured entries with customizable sections and tags
- Keyword and token search for efficient recall
- Non-destructive versioning with diff and restore
- Soft delete and duplicate consolidation
- 60 bundled agent skills accessible via web UI and MCP
- MCP integration (stdio and SSE) for IDE and client use
- Web UI for supervision, statistics, and settings
Use cases of Rememb
- Resume an agent session with full context from prior work
- Maintain an auditable trail of decisions made by AI agents
- Reduce re-explaining project facts (tech stack, architecture) each session
- Quickly search for specific context without rereading entire chat history
- Store user preferences and recurring requests for consistent agent behavior
FAQ from Rememb
What makes Rememb different from cloud-based memory services?
Rememb is local-first: it stores memory as plain JSON or SQLite on disk, with no servers, API keys, or accounts. You keep full control and can copy the store anywhere.
What are the runtime dependencies?
Rememb requires Python 3.10 to 3.12. No additional embedding models or external services are needed at runtime.
Where does Rememb store data?
Data is stored locally in the ~/.rememb/ directory by default. You can copy this folder to move the memory store to another machine.
Which transports does Rememb support for MCP?
Rememb supports stdio (default for IDE integration) and SSE for a persistent local process. The SSE mode must be started separately with the --transport sse flag.
Does Rememb require authentication?
No. Rememb operates entirely locally and does not use any authentication or API keys. Access is limited to the local machine.
More Developer Tools MCP servers
test
cloudwegoThe ultimate LLM/AI application development framework in Go.
MCP-Scan: An MCP Security Scanner
invariantlabs-aiSecurity scanner for AI agents, MCP servers and agent skills.
TalkToFigma
sonnylazuardiTalkToFigma: MCP integration between AI Agent (Cursor, Claude Code, Codex) and Figma, allowing Agentic AI to communicate with Figma for reading designs and modifying them programmatically.

Sentry
modelcontextprotocolModel Context Protocol Servers
FastAPI-MCP
tadata-orgExpose your FastAPI endpoints as Model Context Protocol (MCP) tools, with Auth!
Comments