Skip to content

Commit

Byte's commit system uses AI to generate meaningful commit messages by analyzing your staged changes. It automates the entire commit workflow - staging files, running linters, generating commit messages, and handling git operations.

Commit process showing staging, linting, and message generation

Quick Start

Create a commit:

> /commit

Byte will:

  1. Stage all changed files (with confirmation)
  2. Run configured linters on changed files
  3. Prompt you to choose between Commit Plan or Single Commit
  4. Analyze the git diff with AI
  5. Generate commit message(s) based on your choice
  6. Create the commit(s)

Commit Types

Byte offers two approaches to organizing your commits:

Commit Plan

Analyzes your staged changes and intelligently groups related files into multiple logical commits. The Agent:

  • Reviews all staged files and their diffs
  • Groups related changes together (e.g., feature additions, bug fixes, refactoring)
  • Generates a separate commit message for each group
  • Creates multiple commits automatically

When to use:

  • Large changesets with multiple logical changes
  • Mixed feature work and bug fixes
  • Changes spanning multiple domains or modules

Single Commit

Creates one commit for all staged changes with a single descriptive message. The AI:

  • Reviews all staged files and their diffs
  • Generates one comprehensive commit message
  • Creates a single commit

When to use:

  • Focused changes to a single feature or fix
  • Small, cohesive changesets
  • When you want manual control over commit boundaries

Error Handling

If a commit fails (for example, when pre-commit hooks modify files), Byte will:

  • Display the error message
  • Prompt you to stage the changes and try again
  • Re-stage all changes if you confirm
  • Retry the commit with the same message

This is common with pre-commit hooks that automatically format or modify files - the hooks run, change the files, and cause the commit to fail because the staging area no longer matches the working directory.

Configuration

Linting Integration

Control linting behavior via .byte/config.yaml:

lint:
  enable: true # Disable to skip linting during commits
  commands:
    - command: "ruff format"
      extensions: [".py"]

See Linting concept for detailed configuration.

Commit Message Format

The AI generates messages following conventional commit conventions for both commit types:

Structure

<type>(<scope>): <subject>

<body>

<footer>

Types

Common commit types the AI uses:

  • feat - New feature
  • fix - Bug fix
  • docs - Documentation changes
  • style - Code style changes (formatting, whitespace)
  • refactor - Code restructuring without behavior change
  • perf - Performance improvements
  • test - Test additions or modifications
  • chore - Build process, tooling, dependencies

Examples

Simple feature:

feat: add password reset functionality

Bug fix with scope:

fix(auth): prevent token expiration race condition

Breaking change:

feat!: redesign authentication API

BREAKING CHANGE: AuthService.login() now returns a Promise
instead of synchronous result. Update all callers to use await.