Aider MCP Server

Example Use Cases

• 1

  Generate, fix, or enhance code in a repository via programmatic editing calls (ai_code) using configurable editable/read-only file lists.

• 2

  Discover and choose appropriate LLM models available to the Aider environment (get_models) before performing tasks.

• 3

  Ask general questions or fetch short natural-language responses from an LLM (ask_question) integrated into an MCP workflow.

Description

This server wraps the Aider AI coding assistant to provide standardized MCP tools for one-shot code generation, edits, and question answering. It supports configurable sessions, model selection across providers (OpenAI, Anthropic, Google), and multiple transports (SSE or stdio) for flexible integration. By orchestrating multiple focused LLM calls, it aims to produce more reliable code and reduce costs compared to single monolithic calls. It's intended for embedding AI coding workflows into MCP-enabled clients and agent systems.