WebChatGPT - augment your prompts to ChatGPT with web search results

Intercepts user prompts sent to ChatGPT and automatically enriches them with current web search results before submission. The extension queries a search API (likely Google Custom Search or similar), retrieves top results, and injects formatted search snippets into the prompt context, enabling ChatGPT to reference real-time information beyond its training cutoff. This works by hooking into the ChatGPT UI's message submission flow and prepending search results to the user's original query.

Allows users to select and configure which web search API provider to use (Google Custom Search, Bing Search, DuckDuckGo, or others) through extension settings. The extension abstracts the search provider interface, handling authentication, query formatting, and result parsing for multiple backends. Users can switch providers without code changes by updating extension configuration, enabling flexibility for different rate limits, privacy preferences, or API costs.

Transforms raw search API responses (JSON, XML, or HTML snippets) into a structured, human-readable format that is prepended to the user's original prompt before submission to ChatGPT. The extension parses search results to extract title, URL, and snippet, then formats them as markdown or plain text that ChatGPT can easily consume. This formatting ensures ChatGPT understands the source of information and can cite results accurately in its response.

Manages the extension's runtime lifecycle (initialization, message passing, content script injection) and integrates with ChatGPT's DOM to detect user input, intercept form submission, and inject augmented prompts. The extension uses content scripts to hook into the ChatGPT web interface, listening for user interactions and modifying the DOM before the prompt is sent to OpenAI's API. This requires careful timing to avoid race conditions and ensure the augmented prompt is submitted atomically.

Provides settings to customize how the user's prompt is transformed into a search query, including options to modify query length, add/remove keywords, filter by date range, or exclude certain domains. Users can define custom rules or templates that transform their ChatGPT prompt into an optimized search query before it's sent to the search API. This enables fine-tuning of search results without changing the original prompt to ChatGPT.

Caches search results for identical or similar queries within a session or across sessions (depending on configuration) to reduce API calls and improve response latency. The extension implements a simple cache key based on the search query, storing results in browser local storage or memory. When a user submits a similar prompt, the extension checks the cache before making a new API call, returning cached results if available. Deduplication logic removes duplicate results from the same or different sources.