🚀 Wappler Server Connect — OpenAI ChatKit

Integrate OpenAI's ChatKit into your Wappler applications with Server Connect actions.

This module provides three Server Connect actions for creating chat sessions, sending messages, and retrieving conversation threads.

🔑 Required environment variables

• OPENAI_API_KEY — Your OpenAI API key

Tip: For production, store this in your config.json / user_config.json.

🧰 Wappler Server Connect actions

1) 🤖 Create ChatKit Session

Creates a new ChatKit session that returns a client secret for frontend integration.

Configuration options:

• Session settings (TTL, user ID, metadata)
• Chat configuration (model, instructions, temperature, max tokens)
• Tools configuration (code interpreter, file search)

2) 💬 Send ChatKit Message

Sends a message to a ChatKit session and gets the AI response.

Configuration options:

• Session ID (required)
• Thread ID (optional, for continuing conversations)
• Message content (required)
• File attachments (optional)

3) 📋 Get ChatKit Thread

Retrieves a conversation thread with its messages.

Configuration options:

• Session ID (required)
• Thread ID (required)
• Include messages option

🔄 Usage Flow

1. Create a Session - Use "Create ChatKit Session" action to initialize
2. Send Messages - Use "Send ChatKit Message" action for conversation
3. Retrieve History - Use "Get ChatKit Thread" action to get conversation history

🎨 Frontend Integration

After creating a session on the server, use the returned client_secret in your frontend ChatKit component.

For detailed frontend integration instructions, see the OpenAI ChatKit documentation.

📚 Resources

• OpenAI ChatKit Documentation
• ChatKit Starter App
• Advanced Samples

Troubleshooting

Error: "OPENAI_API_KEY environment variable is not set"

• Make sure you've set the OPENAI_API_KEY in your config.json or environment variables

Error: "session_id is required"

• Ensure you're passing the session_id from the Create Session action

Error: "Run failed with status: failed"

• Check your OpenAI API quotas and rate limits
• Verify your model is available (gpt-4o, gpt-4-turbo, etc.)
• Review the system instructions for any issues

Error: "Request timeout"

• The assistant took too long to respond (over 60 seconds)
• Try reducing the complexity of your request or instructions

📄 License

MIT License - See LICENSE file for details