Automate ChatGPT in your browser. Type → send → wait → run your code. Batch‑friendly. Zero copy‑paste.
Why this exists — Prompting is easy. Repeating it 100× with clean logs, templating, and API glue isn’t. This gives you an on‑page automation panel that behaves like a tiny orchestration tool.
- On‑page automation panel — compose multi‑step chains (Prompt → JS → HTTP → Prompt).
- Batch mode — JSON arrays, objects, or generator functions.
{item},{index},{steps.id.response}supported. - JS sandbox — post‑reply JavaScript with
response,steps,http,utils.log. - HTTP proxy — background service worker makes cross‑origin calls (GM‑style).
- Image capture — auto‑extract generated image URLs.
- Theme‑aware UI — dark/light, draggable, resizable, persistent state.
New! Firefox addon: https://addons.mozilla.org/en-US/firefox/addon/chatgpt-automation-pro/
- Install Tampermonkey
- Open the raw script URL and accept install:
- Go to chatgpt.com → click “Automation” in the header
- Types and sends your message automatically
- Waits for ChatGPT to finish, then runs your JavaScript
- Batch mode with templates: {item.foo}, {{index}}, nested paths
- Polished panel: tabs, progress, logs, dark mode, saved position
- CORS-safe HTTP helper to call any API from your code
- Auto-detects your ChatGPT interface language via
<html lang>and applies built-in translations to the automation panel (supports 30 languages: Albanian, Amharic, Arabic, Armenian, Bengali, Bosnian, Bulgarian, Burmese, Catalan, Chinese, Croatian, Czech, Danish, Dutch, Estonian, Finnish, French, Georgian, German, Greek, Gujarati, Hindi, Hungarian, Icelandic, Indonesian, Italian, Japanese, Kannada, Kazakh, Korean). All strings are bundled locally—no external translation services. - Every panel label, help tip, and log message is translated, so the interface is fully localized in your language of choice.
- Simple tab: paste a message → Send
- Template tab: add Dynamic Elements (JSON array) and a message template → Send
- Response (JS) tab: paste code that runs after the reply
Context available to your JS: response, log, console, item, index, total, http
- Run. Watch it process all cities.
- Chain steps:
prompt,template(batch),js,http. - Templating:
{item},{index},{total},{steps.stepId.response},{steps.httpId.data}. - Context for JS:
response, item, index, total, steps, http, utils.log.
Example: API → JS → Prompt
{
"dynamicElements": ["Tokyo","London"],
"entryId": "weather",
"steps": [
{ "id": "weather", "type": "http", "url": "https://wttr.in/{item}?format=j1", "method": "GET", "next": "extract" },
{ "id": "extract", "type": "js", "code": "const d=JSON.parse(steps.weather.rawText); return d.current_condition[0].temp_C + '°C';", "next": "chat" },
{ "id": "chat", "type": "prompt", "template": "In {item} it\u2019s {steps.extract.response}. What should I wear?" }
]
}Issues and PRs welcome. If this saves you time, ⭐️ star the repo — it helps others find it!