Let the AI Always Work, Then Check: Seamless Task Queuing for Busy Minds.

Sometimes you have multiple, repetitive tasks for ChatGPT, Gemini, Grok, or Claude—like asking them to draft variations, generate batches of images, or produce several code snippets. Manually pasting a prompt, waiting, then pasting the next is slow and distracting. This extension lets you queue all prompts, hit Start, and get back to work while it runs for you.

Auto-Prompt automates sending a sequence of prompts to AI chat sites (ChatGPT, Gemini, Grok, Claude). Paste your list, click Start, and it will:

• Send the first prompt
• Wait until the AI finishes
• Send the next prompt
• Repeat until done

You can leave the page or close the popup. The background worker keeps running.

If you have more ideas to improve efficiency with these AI products, open an issue or ping me—happy to iterate.

• ChatGPT: chat.openai.com, chatgpt.com
• Google Gemini: gemini.google.com
• Grok: grok.x.ai
• Claude: claude.ai

1. Clone this repo
2. Open Chrome → chrome://extensions
3. Enable Developer mode
4. Click Load unpacked → select the repo folder

• manifest.json: MV3 config (permissions, background, popup)
• background.js: Orchestrates prompts, persists settings/history, sends notifications
• content.js: Injected into pages; fills inputs, clicks send, waits for completion
• popup.html + popup.js: UI for prompts, settings, history, theme; progress bar

1. Open one of the supported sites and log in
2. Click the extension icon to open the popup
3. Paste prompts (one per line)
4. Optionally set:
   • Max wait (seconds)
   • Stable (seconds)
   • Poll interval (seconds)
   • System prompt and whether to prepend it to each prompt
   • Theme (Dark/Light)
5. Click Start Automation
6. Watch progress and status; a notification appears when done

• Click Save current as history to store the current prompts + settings
• The list shows recent histories (deduplicated)
• Use Load to restore a specific history
• Use Delete to remove entries

• UIs of AI sites change frequently. If detection breaks, update selectors in content.js
• ChatGPT uses a ProseMirror editor; we insert text using editor-safe events
• Completion detection:
  • ChatGPT: send-button readiness + DOM stability + absence of stop
  • Gemini/Grok/Claude: absence of stop/spinner/typing indicators
• You can tweak default timing in the popup; seconds are converted to ms internally

• Edit files in place and reload the extension from chrome://extensions
• Use DevTools for background (service worker) and popup to debug

MIT