Revathryn

Understand your codebase. Discover what needs attention.

Revathryn is an AI-assisted code review platform that analyses software repositories, identifies potential issues and helps developers understand where their code needs attention.

View on GitHub Explore the roadmap

Focused analysis, structured findings

Revathryn analyses software repositories using specialised review engines. These engines inspect code, identify potential issues and produce structured findings that developers can investigate.

Repository understanding

Discover repository structure, files and projects so analysis operates on a solid foundation.

Structured findings

Analysis outcomes are consolidated into actionable issues with severity, context and supporting reasoning.

Engineering insight

The goal is useful engineering insight, not simply a large volume of AI commentary.

Key capabilities

Revathryn is built around a set of complementary capabilities. The roadmap distinguishes what is implemented today from what is planned.

Repository analysis

Implemented

Scan repositories, discover source files and track file fingerprints to support incremental analysis.

  • Repository scanning
  • Source discovery
  • File fingerprints
  • Incremental analysis

AI-assisted code review

Implemented

Analyse source code, identify potential defects and design concerns, and produce structured issues with supporting context.

  • Source analysis
  • Defect identification
  • Design concerns
  • Supporting context

Specialised review engines

Implemented

Different review types for different engineering concerns, with language-specific analysis and separated responsibilities.

  • Review types
  • Language-specific analysis
  • Separated responsibilities

Issue management

Implemented

Browse findings, review severity and priority, and inspect the reasoning and context behind each finding.

  • Browse findings
  • Severity and priority
  • Reasoning and context

Independent AI validation

Implemented

A separate review process validates previously identified issues to add confidence before accepting findings or fixes.

  • Separate validation
  • Re-check findings
  • Additional confidence

More to come

Planned

Additional review types, language analysers and validation capabilities are planned for future phases.

View the roadmap

How it works

From repository to actionable findings in a repeatable pipeline.

  1. Repository A repository is registered and made available for analysis.
  2. Discovery Source files are discovered and fingerprinted for efficient change detection.
  3. Review Engines Specialised engines inspect code for defects and design concerns.
  4. Issues Findings are consolidated into structured issues with context.
  5. Developer Investigation Developers review, investigate and resolve findings.

Open source

Revathryn is an open-source project. Explore the source code, report bugs, suggest improvements and contribute to development.

View on GitHub