nautex
MCP server for guiding Coding Agents via end-to-end requirements to implementation plan pipeline
claude mcp add --transport stdio hmldns-nautex uvx nautex mcp
How to use
This MCP server integrates Nautex AI with Coding Agents to orchestrate guidance, tasks, and file references for coding projects. Nautex acts as an architect, product manager, and project manager, pulling guidance instructions from Nautex and translating them into structured MCP tasks for the agents. The server exposes instructions and to-do items that reference affected files and requirements, enabling agents to operate in small, well-scoped steps and to map progress to a mapped project structure. Supported agents include Claude Code, Codex, Cursor, OpenCode, and Gemini CLI, each of which can connect to Nautex MCP to receive tasks, status updates, and guidance. To use it, configure the MCP entry for Nautex in your preferred agent tooling (Cursor, Claude Code, Codex, OpenCode, or Git-based integrations), then trigger a test run or a full plan execution to pull the latest Nautex rules and proceed with the next scope.
How to install
Prerequisites:
- A supported environment (macOS, Linux, or Windows) with a functioning Node.js/Python toolchain, depending on your MCP client.
- Access to Nautex AI (sign up and obtain an API token as needed by your workflow).
- Internet connectivity to install and run the Nautex MCP tooling.
Installation steps:
- Install uv (if not already installed). The project demonstrates using uv as the MCP runner. Example:
- macOS/Linux: curl -LsSf https://astral.sh/uv/install.sh | sh
- Windows: powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
- Install and set up Nautex MCP integration by following Nautex onboarding for MCP usage (sign in to Nautex, create API token, and configure the MCP in your agent tools).
- Ensure you have a valid MCP configuration in your agent tooling. An example configuration snippet for Cursor is provided in the Setup details, and similar integration scaffolding exists for Claude Code, Codex, and OpenCode. You can merge Nautex MCP settings into the agent configurations as demonstrated in the README.
- Validate the MCP setup by testing the connection and retrieving the next scope with: uvx nautex mcp test next_scope
- Start using the MCP by pulling rules and proceeding to the next scope once the server reports readiness (status).
Additional notes
Tips and common considerations:
- Ensure your Nautex API token is kept secure and is accessible to the MCP integration you are using.
- If you use multiple tool integrations (Cursor, Claude Code, Codex, OpenCode), keep their MCP configurations in sync. The README shows agent-specific configuration examples and file locations (e.g., .cursor/mcp.json, opencode.json).
- To verify connectivity, you can run uvx nautex mcp test next_scope and check the MCP status via the respective agent tooling UI or CLI.
- When configuring in OpenCode, Nautex writes/updates opencode.json with a minimal required structure; ensure backups are enabled where applicable.
- Use the “Pull nautex rules and proceed to the next scope” workflow in your agent UI to maintain alignment with Nautex guidance.
Related MCP Servers
ReActMCP
ReActMCP is a reactive MCP client that empowers AI assistants to instantly respond with real-time, Markdown-formatted web search insights powered by the Exa API.
mcp-yfinance
Real-time stock API with Python, MCP server example, yfinance stock analysis dashboard
pfsense
pfSense MCP Server enables security administrators to manage their pfSense firewalls using natural language through AI assistants like Claude Desktop. Simply ask "Show me blocked IPs" or "Run a PCI compliance check" instead of navigating complex interfaces. Supports REST/XML-RPC/SSH connections, and includes built-in complian
cloudwatch-logs
MCP server from serkanh/cloudwatch-logs-mcp
servicenow-api
ServiceNow MCP Server and API Wrapper
the -company
TheMCPCompany: Creating General-purpose Agents with Task-specific Tools