Selvage FAQs

Question 1

Can Selvage handle large code changes or extensive projects?

Accepted Answer

Yes, Selvage is optimized for large-scale code reviews. It features Tree-sitter based AST for minimal, relevant context extraction and automatic multi-turn processing to handle reviews exceeding typical LLM context limits stably.

Question 2

Do I need an API key to use Selvage?

Accepted Answer

While Selvage can use external AI model APIs, it also supports an 'Agent-Delegated Review' mode. In this mode, it provides structured review context (diff + Smart Context) to host agents (like Claude Code) which then perform the review using their own LLM, thus requiring no external API key from Selvage users.

Question 3

Which AI models does Selvage support for code reviews?

Accepted Answer

Selvage offers extensive support for multiple leading AI models, including OpenAI GPT (e.g., GPT-5), Anthropic Claude (e.g., Sonnet-4), and Google Gemini, allowing users to choose their preferred LLM.

Question 4

What is Selvage?

Accepted Answer

Selvage is an AI-powered CLI tool designed to automate code reviews. It analyzes Git diffs to help developers improve code quality, detect bugs, and identify security vulnerabilities efficiently.

Question 5

How does Selvage integrate into a developer's workflow?

Accepted Answer

Selvage seamlessly integrates with Git workflows, allowing reviews of staged, unstaged, or commit/branch changes. It also supports MCP Mode and plugins for IDEs like Cursor and Claude Code, enabling natural language review requests.

Selvage

Selvage

主な機能

ユースケース

主な機能

ユースケース