// devtools · repowise-dev/repowise

repowise

repowise ist ein Open-Source-Projekt aus der Kategorie Entwickler-Tools von repowise-dev (repowise-dev/repowise auf GitHub). Es hat 7.239 Sterne auf GitHub und 791 Forks, wird unter der Lizenz AGPL-3.0 veröffentlicht und wurde zuletzt am 8. Oktober 2026 auf GitHub aktualisiert.

Repowise — codebase intelligence: health scores, auto-docs, git analytics and dead-code detection via MCP.

★ 7,2 Tsd. Sterne⑂ 791 ForksAGPL-3.0aktualisiert Okt 2026
▦ Visual — das README dieses Projekts enthält keinen einbettbaren Screenshot,
darum hier ein Platzhalter. Die Anleitung unten ist trotzdem vollständig.
// projektdaten

Projektdaten

GitHub-Sterne
7.239
Forks
791
Lizenz
AGPL-3.0
Zuletzt auf GitHub aktualisiert
8. Oktober 2026
Eintragstyp
Vollständige Installationsanleitung
Offizielle Website
Official site
// so funktioniert es

So funktioniert es

  1. Clarify your goal

    Pick the repo to index and decide which front door you will use — CLI for you, MCP for your agents, or both.

  2. Get the project

    Clone repowise-dev/repowise from GitHub (or use the install path in the run steps below) and skim its README for the parts relevant to your goal.

  3. Install and configure

    Install repowise as a uv tool (or pipx/pip), then run its one-line setup in your repository to build the index.

  4. Run your first task

    Index the repo, read its health score, then ask your agent (via MCP) which files are risky to change in your next task.

// funktionen

Warum es so beliebt ist

Code health scores

Scores repository health and re-checks its own score after each index, so quality trends are visible, not vibes.

Auto-generated docs and agent files

Generates documentation and writes your CLAUDE.md and AGENTS.md from the real index, keeping them current.

Change-risk scoring

Scores any commit or base..HEAD range 0–10 from the shape of the change, plus bug-history per file and symbol.

Test intelligence

Works out which tests reach a file and which a diff exercises, and estimates test coverage without running coverage.

Dead code and git analytics

Dead-code detection, architectural-decision records, and analytics over git history, including which other open branches edit your files.

Agent-ready via MCP

Exposes the index through MCP, distills large command output before agents read it, and seeds new git worktrees with the index automatically.

Advertisement
// selbst ausführen

Führe es selbst aus

  1. Clone the repository

    Get the source from the official GitHub repository for repowise.

    git clone https://github.com/repowise-dev/repowise.git
    cd repowise
  2. Install Repowise

    Install it as a uv tool exactly as the README's one-line setup shows (pipx and pip installs are also documented).

    uv tool install repowise
  3. Set up your repository

    Run Repowise's setup/index in your repo so the .repowise/ index is built next to the code.

    # Run the repowise setup command in your repository per the README — see README for the exact command
// anwendungsfälle

Was du damit bauen kannst

Advertisement
// häufige fragen

Häufige Fragen

Who maintains repowise?

repowise-dev, through the repowise-dev/repowise repository, with the site at repowise.dev.

Is repowise free to use?

Yes — the project is open source under the AGPL-3.0 license as recorded for this listing. See the license terms for any conditions on redistribution.

Does repowise need a paid API key?

The self-hosted open-source version runs on your machine with your own LLM provider only if you turn its prose features on; core indexing does not require a paid key.

How does Repowise compare to a linter or static-analysis tool?

Linters check code style and known bug patterns file by file. Repowise builds a repo-wide intelligence layer — health, history, risk, tests, dead code and decisions — and serves it to agents via MCP, which is context for decisions, not just diagnostics.

How do I get started with repowise?

Clone the repository (or use the documented installer), follow the install and configure steps on this page, and run the smallest real task first — index the repo, read its health score, then ask your agent (via MCP) which files are risky to change in your next task.

// weiter entdecken

Ähnliche Tools

Entwickler-Tools

fallow

Codebase intelligence for TypeScript/JavaScript — health, duplication, dead code and architecture.

★ 5,0 Tsd. SterneMIT
Anleitung ansehen →
Entwickler-Tools

code-review-graph

Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked con…

★ 31,9 Tsd. SterneMIT
Anleitung ansehen →
Entwickler-Tools

opcode

A desktop GUI and toolkit for Claude Code - agents, sessions, analytics and MCP

★ 22,4 Tsd. SterneAGPL-3.0
Anleitung ansehen →
Entwickler-Tools

claude-context

Code-search MCP for Claude Code: make your entire codebase the context for any coding agent.

★ 12,6 Tsd. SterneMIT
Anleitung ansehen →
Entwickler-Tools

fastapi_mcp

Expose your FastAPI endpoints as Model Context Protocol tools - with authentication - in a few lines of code.

★ 12,0 Tsd. SterneMIT
Anleitung ansehen →
Entwickler-Tools

ralph-claude-code

An autonomous loop for Claude Code with intelligent exit detection, rate limits and a circuit breaker.

★ 9,7 Tsd. SterneMIT
Anleitung ansehen →
// alternativen

Alternativen im Bereich Entwickler-Tools

Die anderen Projekte der Kategorie Entwickler-Tools mit den meisten Sternen im LibreAI-Katalog — die realistische Auswahl, falls repowise nicht das Richtige ist.

// mehr in devtools

Mehr in Entwickler-Tools

Verwandte Projekte der Kategorie Entwickler-Tools im Katalog, die repowise bei den Sternen am nächsten stehen — zum Weiterstöbern auf beiden Seiten.

Hol dir repowise

Kostenlos und offen — hol dir den Quellcode, gib dem Repository einen Stern und leg los.