Overview
- Predict which files will break next with 73% accuracy using code health scoring that weighs tangled complexity, missing tests, churn, and hidden coupling
- Catch risky commits and pull requests before merge with defect-risk scores for every commit and PR
- Give Claude Code, Cursor, Codex, and any MCP client real codebase context through 9 integrated MCP tools covering structure, dependencies, and architecture
- Eliminate stale documentation with an auto-wiki that generates and continuously updates a complete codebase wiki without manual effort
- Untangle hidden dependencies and bus factor risks using Git intelligence that flags concealed coupling, hotspots, and sensitive code from Git history
- Know exactly who owns every file with Git blame-powered ownership insights showing primary owners and top 3 contributors per code section
- Cut through CVE noise with reachability-aware security triage that scans your dependency graph to identify whether vulnerable code is actually used
- Measure AI-written code quality and quantity with agent provenance that tracks and evaluates code health and ownership of AI contributions
- Query your repo's architecture, dependencies, and structure directly from Claude Code, Cursor, or any MCP client using natural language
- Get a complete architectural view with C4 system context, containers, and components extracted from inline markers, git archaeology, README mining, and CLI
- Coordinate entire teams from one shared index with a single credit pool and organizational installation, eliminating redundant repository rediscovery
- Deploy on your terms with self-hosted, air-gapped, and commercially licensed options for strict security and compliance requirements
Pros & Cons
Pros
- Detailed code health insights
- Agent provenance analysis
- Change risk assessments
- Security Reachability-aware feature
- CVE triage
- Automated documentation with auto-wiki
- Git Intelligence Analysis
- C4 system architectural insights
- Architectural decision extraction
- Repo model querying capability
- Data centralization
- Risk detections with flags
- Ownership determination
- Hidden coupling detection
- Offers self-hosted options
- Commercial licenses available
- Air-gapped hosting option
- Repo analysis capability
- Team coordination facilitation
- Fully indexed dependency graph
- Auto-generated wikis
- Dependency graphs generation
- Hotspot identification
- Supports multiple MCP clients
- One organizational installation
- File health tracking
- Critical symbol identification
- Real-time architecture overview generation
- Extensive doc generation
- Defect risk predictions
- Effort-for-impact prioritization
- Staleness tracking for decisions
- Co-change partner identification
- CLI, MCP server, and Web UI
- Integrates with GitHub App
- Free for public/OSS repos
- Retrieves over wiki gates
- Natural language query support
- Semantic search over full wiki
- Includes editor intelligence
- Automates Claude.md file
- Dead code identification and summary
- Supports private repos
- Auto-generation of documentation
- Tree-sitter and NetworkX backed
- Ranks refactoring targets
- Raw source byte retrieval
- Health degradation detection
- Supports various editors
- Dead code detection
- Supports tech stack identification
Cons
- Requires specialized MCP clients
- Centralized data may limit privacy
- No indication of cost
- Interface might be complex
- Requires extensive codebase
- CVE Triaging might be slow
- Versatile, potentially overwhelming interface
- Lack of direct support
- Prone to false positives
Reviews
Rate this tool
Loading reviews...
❓ Frequently Asked Questions
Repowise is an intelligence tool that provides in-depth insights into a codebase. It offers a variety of features including code health, agent provenance, AI context, change risk, security, auto-wiki, Git intelligence, architecture components and more. It also integrates the capability to query models of your repo using tools such as Claude Code and Cursor and evaluates the health and ownership of code written by AI. Repowise supports self-hosted, air-gapped and commercially licensed options.
Repowise's code health feature provides a verified score for each file in your repository. The health score is created based on factors such as tangled complexity, hidden coupling, missing tests, and churn. Its scoring system aims to predict the probability of a file encountering a bug in the future, with a state-of-the-art accuracy of about 73%. This means it can give a heads-up on which file is probable to encounter a bug next before it does, offering insights on what needs to be fixed first based on impact-for-effort.
Agent provenance in Repowise refers to the analysis of the proportion of your code written by AI. This feature scrutinizes and assesses the quantity and potentially, the quality of the code produced by AI entities within your project.
Repowise utilizes AI context and MCP (Mental Model of Codebase Prior) tools to furnish your agent with real codebase context. It does this by integrating 9 MCP tools that give your AI agent the information it needs to understand the structure and dependencies within your codebase. Repowise allows developers to query models of your repo using Claude Code, Cursor, or any other MCP client.
The change risk feature in Repowise delivers a defect-risk score for each commit or Pull Request (PR). This score assists in identifying potential problems in commits or PRs and helps mitigate future code issues by flagging risky PRs.
The Security Reachability-aware feature of Repowise allows for CVE (Common Vulnerabilities and Exposures) triage on your dependency graph. This means that it identifies the usage of vulnerable code in your project and helps in making informed decisions on fixing or replacing it.
The auto-wiki feature in Repowise is designed to create a fully documented wiki of your codebase that can update itself. This generated wiki covers all dimensions of your codebase, resulting in a comprehensive document that requires no manual updating.
The Git intelligence feature of Repowise pinpoints hotspots, ownership, concealed coupling, and bus factor in your codebase. It sifts through your Git history and uses complex metrics to identify areas of code which are particularly sensitive or prone to cause problems, flagging them for review.
Repowise integrates an architecture component that provides C4 system context, containers and components. This assists in gaining a broad architectural view of your codebase, giving you a comprehensive overview of your system's structure and interactions.
Repowise extracts architectural decisions from a multitude of sources. This feature assembles an in-depth view of the architecture by pulling information from resources such as inline markers, git archaeology, README mining, and Command Line Interface (CLI) among others.
Yes, Repowise allows developers to query models of your repo using Claude Code, Cursor, or any other MCP client. It takes these queries and returns information about the structure, dependencies, and other aspects of your codebase.
Repowise centralizes data in one shared index. It creates a single point of access for all your codebase data, facilitating efficient coordination for teams. It does this with one credit pool and one organizational installation, making information quickly and efficiently accessible.
For engineers, Repowise evaluates the code written by AI, determining the code health and ownership. This allows for a wrapping understanding of Code Quality and agent contribution to the system.
Yes, Repowise supports self-hosted, air-gapped, and commercially licensed options. The flexibility of deployment supports discretion across different business standards and sizes.
Repowise assists with CVE (Common Vulnerabilities and Exposures) triage by identifying the usage of vulnerable code. The Security Reachability-aware feature in Repowise scans your dependency graph for potential security issues, providing a heads-up on the vulnerabilities and thereby assisting with their resolution.
Repowise performs repository analysis by employing its catalogue of features such as code health, AI context, agent provenance, change risk, security, and Git intelligence among others. With these features, it provides a fine-grained overview and allows developers to query models of their repository, leading to a comprehensive and deep-diving analysis.
The feature of Team Coordination in Repowise centralizes data in one shared index. This improves team cooperation by ensuring all members access the same, most up-to-date data from one point, be it from one credit pool or one organizational installation.
Repowise detects hidden coupling through the Git Intelligence feature. It flags such instances, enabling engineers to untangle dependencies and streamline their codebase.
Repowise determines ownership based on the Git intelligence feature, which gives insights into who made what changes in the code from the Git blame data. It can identify the primary owner and the top 3 contributors for each piece of code.
The Codebase Documentation feature of Repowise refers to the comprehensive documents it auto-generates through its auto-wiki feature, covering the entirety of your codebase. It keeps these documents up to date with the ongoing changes in your codebase without any manual intervention required.
Pricing
Pricing model
Freemium
Paid options from
$12/month
Billing frequency
Monthly
