
CodeCartographer turns a repository into layered architecture notes, behavioral contracts, defect findings and a language-agnostic reimplementation spec. It ships as a Pi extension, an MCP server and a drop-in template.
PluginContact for pricing
Step 1: Install the package globally with npm install --global codecartographer-pi.
Step 2: Add a codecartographer server entry to your MCP host config with the command codecarto-mcp.
Step 3: Call codecarto_init with the absolute path of the target repository and choose a pipeline.
Step 4: Use codecarto_next to get the next phase prompt, run it in your agent, then call codecarto_validate and codecarto_complete.
Step 5: Check progress with codecarto_status or regenerate the report with codecarto_dashboard.
The Pi extension is the primary surface. The MCP server works with Claude Code, Claude Desktop or any MCP-compatible host, and a drop-in template works with any LLM that can read and write files.
Each phase must validate against its template before the status advances. Findings carry one of five evidence tags, and an unsettled level cannot carry a settled fix action.
A batch reconnaissance step that runs six analysis lenses over a git repository through the OpenRouter Batch API. Its output is unverified leads that later phases must confirm or dismiss.
The project is MIT licensed. Broad-Side uses your own OpenRouter key and is billed there.
Requirements
CodeIndex MCPcodeindex is an MCP server that indexes a codebase with tree-sitter and answers structural questions such as symbol definitions, callers and change impact. It exposes 16 tools to AI coding agents.
Neruva Agent MCPNeruva is a memory layer for coding agents, delivered as an MCP server backed by a hosted API. Agents recall earlier decisions, log what they do and leave handoff notes between sessions.Repo2Graph MCPRepo2Graph MCP gives coding agents cited, token-bounded answers about a codebase. It indexes calls, imports and git co-change links, then returns the source blocks that answer a question.