Chrome extension to add input history, copy, and counters to ChatGPT

Maintains a local browser storage index of all user-submitted prompts within the current ChatGPT session, enabling up/down arrow key cycling through recent inputs and fuzzy-match autocomplete triggered by the '/' character. Implements DOM event listeners on ChatGPT's input textarea to intercept keystrokes and populate a dropdown menu with matching historical prompts, reducing re-typing of common queries and enabling rapid iteration on prompt variations.

Adds copy buttons to each ChatGPT message (both user prompts and assistant responses) that capture the full text content and preserve formatting (markdown, code blocks, line breaks) when copied to the system clipboard. Implemented via injected UI buttons in the ChatGPT message DOM that trigger native clipboard API (navigator.clipboard.writeText) with the extracted message text.

Provides a sidebar panel or modal that displays available custom GPTs from OpenAI's GPT Store with search, filtering, and one-click launch functionality. The extension fetches GPT metadata (name, description, creator, rating) from OpenAI's GPT Store API or scrapes the store interface, displays results in a searchable list, and allows users to launch custom GPTs directly from the extension without navigating to the store. Stores user's favorite/tested GPTs for quick access.

Provides an enhanced sidebar panel for ChatGPT's voice mode that displays the conversation in a side-by-side layout, showing real-time transcription of user speech and ChatGPT's responses. Implements voice input/output handling via Web Speech API (speech recognition and text-to-speech) or integration with ChatGPT's native voice mode, displaying the conversation in a scrollable sidebar while the main ChatGPT interface remains accessible. Allows users to switch between text and voice input seamlessly.

Displays estimated token counts and usage statistics (characters, words, estimated API cost) for each ChatGPT message pair (user input + assistant response). Calculates token estimates using a client-side tokenizer (likely tiktoken.js or similar) that approximates OpenAI's token counting without requiring API calls, injecting counter badges into the ChatGPT UI next to each message.

Provides a sidebar panel that mirrors ChatGPT's conversation list with additional organizational features: drag-and-drop folder creation, color-coded labels, custom tags, and pinned/favorite conversations. Stores folder structure and metadata in browser local storage (or cloud sync in Pro tier) and intercepts ChatGPT's conversation navigation to apply custom sorting/filtering based on user-defined categories. Implements a tree-view UI component with drag-and-drop handlers for moving conversations between folders.

Adds a right-click context menu option to capture the current browser page (or selected region) as a screenshot, automatically inserts the image into ChatGPT's input field, and submits it to ChatGPT's vision model for analysis. Implements screenshot capture via Chrome's tabs.captureVisibleTab API, converts the image to base64, and injects it into ChatGPT's message input using DOM manipulation to trigger the image upload handler.

Provides a library of saved prompt templates with placeholder variables (using {{double curly bracket}} syntax) that can be filled in before submission. Templates are stored in browser local storage or cloud (Pro tier) and accessible via a sidebar panel or slash-command autocomplete. When a template is selected, the extension opens a form to fill in variable values, substitutes them into the template text, and populates ChatGPT's input field with the completed prompt.

Adds a download button to ChatGPT responses that include audio (from ChatGPT's text-to-speech feature), capturing the audio stream and saving it as an MP3 or WAV file to the user's downloads folder. Intercepts the audio element in ChatGPT's DOM, extracts the audio source URL or stream data, and triggers a browser download using the Blob API and URL.createObjectURL.

Provides a 'Optimize Prompt' button (Pro tier feature) that sends the current user input to an AI model (likely ChatGPT itself or a specialized prompt optimization service) to generate improved versions of the prompt with better clarity, specificity, and likelihood of generating high-quality responses. The extension displays optimization suggestions in a modal or sidebar panel, allowing the user to accept, reject, or manually edit suggestions before submitting to ChatGPT.

Adds a model selector dropdown to ChatGPT's interface (or sidebar) that allows switching between available OpenAI models (GPT-4, GPT-3.5-turbo, etc.) without navigating to ChatGPT's settings. The extension intercepts the model selection and either triggers ChatGPT's native model switcher via DOM manipulation or (if using a custom API integration) routes requests to different model endpoints. Stores the user's model preference in local storage for persistence across sessions.

Provides export functionality for entire conversations or individual messages to multiple file formats (PDF, Markdown, JSON, TXT) with optional metadata preservation (timestamps, model used, token counts). Implements format-specific exporters that traverse the conversation DOM, extract message content and metadata, and generate formatted output files. PDF export uses a client-side PDF library (likely jsPDF or similar), while JSON export preserves structured metadata for programmatic access.