Codebase intelligence for AI and humans: code health scores, auto-generated docs, git analytics, dead code detection, and architectural decisions via MCP.
Python aicode-complexitycode-healthcode-intelligencecode-qualitydead-codedeveloper-toolsdocumentationgit-analyticsmcpopen-sourcepythonrefactoringstatic-analysistechnical-debt
Docker
Two images are provided:
Dockerfileruns the full repowise stack (API + Web UI) in a single container.Dockerfile.mcpruns only the MCP server over stdio, for MCP hosts and directories (see MCP server image below).
Run the full repowise stack (API + Web UI) in a single container.
Prerequisites
- Docker installed
- A repository already indexed with
repowise init(the.repowise/directory exists)
Quick Start
# Run from the project root
docker build -t repowise -f docker/Dockerfile .
# Run with a mounted .repowise directory
docker run -p 7337:7337 -p 3000:3000 \
-v /path/to/your-repo/.repowise:/data \
-e GEMINI_API_KEY=your-key \
-e REPOWISE_EMBEDDER=gemini \
repowise
Open http://localhost:3000 for the Web UI, http://localhost:7337 for the API.
Docker Compose
# Set the path to your .repowise directory
export REPOWISE_DATA=/path/to/your-repo/.repowise
export GEMINI_API_KEY=your-key
docker compose up
Environment Variables
| Variable | Default | Description |
|---|---|---|
REPOWISE_DB_URL |
sqlite+aiosqlite:////data/wiki.db |
Database URL |
REPOWISE_EMBEDDER |
mock |
Embedder: gemini, openai, mock |
ANTHROPIC_API_KEY |
— | Anthropic API key (for chat) |
OPENAI_API_KEY |
— | OpenAI API key (for chat) |
GEMINI_API_KEY |
— | Gemini API key (for chat + embeddings) |
PORT_BACKEND |
7337 |
API server port |
PORT_FRONTEND |
3000 |
Web UI port |
MCP server image
Dockerfile.mcp is a lean image that runs only the MCP server, speaking
JSON-RPC over stdio. It carries no Web UI and exposes no ports. This is the
image to point MCP hosts and directories at.
# Build
docker build -t repowise-mcp -f docker/Dockerfile.mcp .
# Run against a repo (mount it and set it as the working dir).
# tools/list works without an index; the query tools serve a mounted .repowise.
docker run --rm -i -v /path/to/your/repo:/repo -w /repo repowise-mcp
The container communicates over stdin/stdout, so run it with -i and no -t.