Readable - AI Generated Comments

Generates multi-line function documentation comments by analyzing the function signature and body when user presses Ctrl+' (Windows/Linux) or Cmd+' (macOS). The extension extracts the function context from the current cursor position, sends it to OpenAI's API via Readable's backend, and inserts the generated docstring at the appropriate location (above the function). Works across JavaScript, TypeScript, Python, C, C#, C++, Java, and PHP by using language-specific AST or regex-based function boundary detection.

Generates single-line comments for code snippets when user types '//' (C-style languages) or '#' (Python) followed by a space, then presses Tab. The extension captures the preceding line(s) of code, optionally incorporates user-typed context words, sends the code snippet to OpenAI, and inserts the generated comment inline. Supports context-aware generation — users can type words after the comment marker to guide the AI toward specific comment types (e.g., '// TODO' or '# warning').

Scans the entire codebase to identify comments that no longer match their associated code (e.g., function documentation that describes outdated parameters or logic). Accessible via a 'Find Stale Comments' sidebar panel, the extension analyzes each comment against its corresponding code block, flags mismatches, and allows users to regenerate comments in bulk. Uses AST or regex-based comment-to-code association to map comments to their targets across all supported languages.

Abstracts away language-specific comment syntax and insertion logic by automatically detecting the language of the current file and inserting generated comments in the correct format and location. Supports 9 languages (JavaScript, TypeScript, JSX/TSX, Python, C, C#, C++, Java, PHP, Rust) with language-specific AST or regex-based parsing to identify function boundaries, class definitions, and appropriate insertion points. Users trigger generation via keyboard shortcut or text trigger without needing to specify language or comment style.

Abstracts OpenAI API key management and rate limiting by routing all comment generation requests through Readable's own backend infrastructure. Users authenticate via GitHub OAuth or email/password on readable.so, and the extension communicates with Readable's API rather than directly with OpenAI. This approach centralizes billing, quota management, and API key security, eliminating the need for users to manage their own OpenAI API keys or worry about exposing credentials in their VSCode configuration.

Implements a freemium model where stale comment detection is available for free, but AI-powered comment generation (docstring, inline, and bulk regeneration) requires a paid subscription ($19.99/year). The extension enforces feature gates at the API level — free tier users can access the sidebar and detection UI but receive errors when attempting to generate comments. This model allows users to evaluate the tool's detection accuracy before committing to paid generation.

Exposes comment generation features via VSCode's command palette with two commands: 'Readable: Enable Comment Suggestions' and 'Readable: Disable Comment Suggestions'. These commands toggle the `readable.enableAutoComplete` setting, allowing users to quickly enable/disable inline comment generation without navigating VSCode settings. Provides an alternative to keyboard shortcuts for users who prefer menu-based workflows or need to disable the feature temporarily.

Allows users to provide optional context words or phrases after the comment marker (e.g., '// TODO' or '# warning') to guide the AI toward specific comment types or tones. The extension captures these user-typed words and includes them in the API request to OpenAI, influencing the generated comment's content and style. This hybrid approach combines the speed of AI generation with user control over comment intent, reducing the need for post-generation editing.

Provides a dedicated sidebar panel in VSCode that displays stale comments detected across the codebase, organized by file and line number. Users can click on stale comments to navigate to them in the editor, and (with paid subscription) regenerate comments in bulk from the sidebar. The panel integrates with VSCode's sidebar UI, allowing users to manage documentation debt without leaving the editor.