This connector is only available for Next.js TypeScript web build tasks.
What you can do
AI chat assistant
Conversational chatbot that answers questions, provides recommendations, or guides users through workflows.
Content generation
Let users generate blog posts, marketing copy, product descriptions, or social media content from simple prompts.
Document summarizer
Upload documents or paste text and get concise summaries, key takeaways, or action items automatically.
AI-powered search
Semantic search that understands user intent instead of just matching keywords.
Customer support bot
Automate first-line support with a bot that answers FAQs, triages issues, and escalates complex requests.
Before you connect
You need an OpenAI API key.Get your OpenAI API key from the dashboard
OpenAI is a task-level connector. Each Rocket task connects to its own API key independently.
Connect OpenAI
- Web Browser
- Mobile App
You can connect from two places - both open the same popup.Option 1: From chatType a prompt that mentions OpenAI or GPT - for example, 

Click the OpenAI card, then click Connect.

After clicking ConnectA popup opens. Paste your API key and click Connect.

A green dot appears next to OpenAI when the connection is active.Update or disconnectOpen Connectors and click the OpenAI card. Click Edit to update the key or Disconnect to remove it from this task.

Connect OpenAI and add a GPT-powered chat assistant to my app. Rocket detects the intent and shows a Connect button inline. Click it and the popup opens.Option 2: From the Connectors tabClick the ... button in the preview toolbar, then select Connectors.







Example prompts
| What you want | Prompt to use |
|---|---|
| Chatbot | Add an AI chatbot to my app that answers user questions using OpenAI GPT. |
| Blog writer | Build a blog post generator where users enter a topic and get a full draft from GPT. |
| Email drafter | Create an email composer that generates professional emails from bullet points using GPT. |
| Code explainer | Add a code explainer tool where users paste code and get a plain-English explanation from OpenAI. |
| Content moderator | Use OpenAI to automatically flag inappropriate user-submitted content before it is published. |
| Quiz generator | Build a quiz generator that creates multiple-choice questions from any topic using GPT. |
| Language translator | Add a translation feature that converts user text between languages using OpenAI. |
| Sentiment analyzer | Analyze customer feedback sentiment using OpenAI and display results in a dashboard. |
| Meeting summarizer | Summarize meeting notes into key decisions and action items using GPT. |
| Product descriptions | Generate product descriptions from basic details like name, category, and features using OpenAI. |
Tips
- API billing is separate. OpenAI charges based on token usage. Monitor your spending in the OpenAI usage dashboard.
- Rate limits apply. New accounts have lower rate limits. Check your rate limit tier and request an increase if needed.
- Token costs vary by model. Longer prompts and responses cost more. Keep system prompts concise and set reasonable max token limits.
- OpenAI vs. Anthropic. GPT models excel at broad general-purpose tasks, function calling, and vision. Claude is better suited for long-context analysis and tasks requiring careful, detailed reasoning.
- One API key per task. Each Rocket task connects to one OpenAI API key independently.
What’s next?
Anthropic
Add Claude models for long-context analysis and detailed reasoning alongside GPT.
Gemini
Connect Google’s Gemini models for multimodal AI features.
Perplexity
Add real-time web search and up-to-date answers to your AI features.
Supabase
Store conversation history, user preferences, and AI-generated content.

