Ctxl Mcp
Give Claude or Cursor durable personal context so coding sessions do not re-explain your stack, preferences, and project facts every time.
Overview
ai.autoblocks/ctxl-mcp is an MCP server for the Build phase that stores and retrieves personal context for AI assistants over stdio.
What is this MCP server?
- stdio npm package ctxl-mcp (v0.0.2) for local MCP clients such as Claude Desktop and Claude Code
- Personal context management scoped to you—not a team wiki or generic RAG pipeline
- GitHub source at autoblocksai/ctxl for self-hosting and fork-friendly customization
- Lightweight stdio transport fits subprocess-spawned MCP without a hosted endpoint
- Complements project rules and CLAUDE.md by storing retrievable personal memory via tools
- Server version 0.0.2 in MCP registry schema
- Single npm stdio package identifier ctxl-mcp
- Repository: github.com/autoblocksai/ctxl
What problem does it solve?
Every new agent session forces you to re-upload the same project background, coding preferences, and constraints.
Who is it for?
Indie developers running Claude Code or Claude Desktop who want a small, local MCP memory layer without standing up a full vector database.
Skip if: Teams that need audited, shared knowledge bases with RBAC—or builders who only need static CLAUDE.md files and never use MCP.
What do I get? / Deliverables
Your MCP client can persist and pull personal context on demand so agent replies stay aligned with how you actually work.
- Registered stdio MCP server exposing personal context tools to your agent
- Reusable personal context store the assistant can query during implementation work
- Local subprocess integration without mandatory hosted API keys from server.json
Recommended MCP Servers
Journey fit
Context layers sit on the agent toolchain shelf because builders wire them into MCP clients during product development, not during idea research. agent-tooling is the canonical home for MCP servers that extend what the assistant remembers and retrieves across repos and sessions.
How it compares
MCP context-memory integration, not a planning skill or static repo documentation file.
Common Questions / FAQ
Who is ai.autoblocks/ctxl-mcp for?
Solo builders and indie developers using MCP-enabled agents who want persistent personal context across coding sessions.
When should I use ai.autoblocks/ctxl-mcp?
Use it during Build (and ongoing Operate) when you are wiring agent-tooling and tired of repeating the same background in every chat.
How do I add ai.autoblocks/ctxl-mcp to my agent?
Add the npm stdio server ctxl-mcp to your MCP client config (Claude Desktop, Claude Code, or compatible hosts), then restart the client so tools load.