AI-assisted “why?” comments for your code changes. WhyComment scans diffs and highlights lines that deserve a short rationale, then helps you insert concise comments where they matter most.
Features
Auto-analyzes saved files (debounced) using a compact Git diff.
“Analyze Selection” to focus on a specific range.
Side panel to review suggestions and jump to lines.
One-click “Apply” to insert a comment above the line, preserving indentation.
“Suggest Comment Variants” to pick from 3 AI-generated one-liners.
Stable line mapping and duplicate suppression for reliable results.
Quick Start
Install the extension from the Marketplace.
Set an LLM API key in Settings (WhyComment: Api Key). Supports Claude (default) and OpenAI.
Save a file or right-click and run “WhyComment: Analyze Selection”.
Open the WhyComment view in the Explorer, review items, and Apply or Ignore.
Commands
WhyComment: Analyze Selection – Analyze the current selection.
WhyComment: Suggest Comment Variants – Generate 3 comment candidates for a line.
WhyComment: Toggle Auto Analyze – Enable/disable auto analysis on save.
WhyComment: Clear All Suggestions – Clear all current suggestions.
Behavior
Debounced auto-analysis (default 1s) on save for workspace files only.
VS Code configs (e.g., .code-workspace, .vscode/*) are ignored by default.
Respects exclude patterns for auto-analysis; manual “Analyze Selection” still runs.
Inserted comments align to the target line’s indentation.
Existing UI items prevent duplicate suggestions for the same line.
Providers & Models
Works with Claude (default) and OpenAI. Choose provider/model in Settings.
The extension sends a compact, annotated diff to the provider to minimize payload.
Privacy & Data
No code is sent unless you configure an API key.
Only compact diffs (and locale) are sent to generate suggestions.
No data is stored by the extension beyond VS Code state for UX.
Requirements
VS Code 1.80+ and a Git-initialized workspace.
An API key for your chosen LLM provider.
Troubleshooting
“No suggestions”: ensure the file has changes vs HEAD, or try Analyze Selection.
“Not analyzing this file”: check exclude patterns and that it’s inside the workspace.
Progress indicator shows during LLM calls. Errors surface as VS Code notifications.
