Skip to main content
MCP server CX Codex

conxt — Claude Code MCP server: what it does, how to install

conxt retains project decisions, coding guidelines, and technical stack details across AI chats. Helps teams keep shared rules in sync.

no license 263 installs updated 2026-06-22

Install into Claude Code · Platform CLI: remote server All methods and platforms ↓
claude mcp add --transport http conxt https://conxt--pgavali0318.run.tools
Open on GitHub

What conxt does

This server connects your coding agent to an external Conxt memory graph via https://conxt--pgavali0318.run.tools. It exposes eight MCP tools for reading pinned entries, pulling task-relevant memories, and recording technical choices. The default context retrieval returns up to 15 entries unless configured otherwise.

Team and individual memories remain completely separated. While add_memory stores personal preferences, add_team_memory pushes guidelines to a shared workspace. Team operations fail without a team UUID, so the model must run list_my_teams beforehand to resolve group IDs. User identification happens automatically through the API key.

What it helps with

  • store architecture choices and coding guidelines across restarts
  • retrieve project context without explaining the technology stack again
  • distribute coding standards to colleagues in shared workspaces
  • search previous technical decisions by specific topic

Example prompts for the agent

  • Check Conxt for our team decisions on database migrations
  • Add a memory that we prefer pnpm over npm for all repository packages
  • List my teams and save this authentication pattern to our shared space

How to install conxt

Platform CLI: remote server

claude mcp add --transport http conxt https://conxt--pgavali0318.run.tools

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 conxt

How to install conxt in Claude Code?

Run claude mcp add --transport http conxt https://conxt--pgavali0318.run.tools. The memory tools will be ready in your next session.

What does conxt do?

It provides long-term memory for coding assistants, capturing project preferences, workflow rules, and decisions. Agents can query this graph whenever starting fresh tasks.

How are team memories managed?

Team data stays isolated from personal notes and requires a team UUID. The agent calls list_my_teams first to look up available team IDs before reading or saving shared records.

Similar MCP servers

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

Project sponsors

Companies that keep this analytics open