tidewave-ai/tidewave_phoenix

What it solves

It provides a way for AI coding agents to interact directly with a running Phoenix (Elixir) application during development. This removes the need for agents to guess how code behaves or rely solely on static analysis by giving them runtime access to the app's state, logs, and database.

How it works

It functions as a Model Context Protocol (MCP) server that integrates into a Phoenix app via a plug. Once installed, it exposes a set of tools that an MCP-compatible agent (like Claude Code or Cursor) can call over HTTP to perform actions inside the live development environment.

Who it’s for

Developers building Phoenix applications who use AI coding agents to assist with development, debugging, and feature implementation.

Highlights

• Runtime Code Execution: Allows agents to evaluate Elixir code in the context of the running app, similar to an IEx session.
• Database Access: Enables agents to execute SQL queries against the development database to verify data state.
• Live Log Access: Lets agents read server logs to diagnose errors and trace request backtraces.
• Version-Aware Documentation: Provides documentation for modules and functions based on the exact versions locked in the project's dependencies.
• Source Mapping: Allows agents to find the exact file and line number of a function definition, including those within external dependencies.