Overview

OpenAI Stream Chat Widget

A powerful custom widget that brings advanced OpenAI capabilities directly into your Mendix applications. This widget goes beyond basic LLM integrations by supporting both Chat Completions and Responses APIs with real-time streaming functionality.

Key Features:

• Dual API Support: Choose between OpenAI's Chat Completions API or the Responses API based on your needs
• Real-time Streaming: Watch AI responses appear character by character with built-in Server-Sent Events (SSE) support
• Flexible Configuration: Pass custom JSON request bodies to easily adjust parameters like reasoning-effort and response-format, without having to set up a domain model or import/export mappings.
• Developer-Friendly: Simple configuration through Mendix Studio Pro with clear parameter inputs for API key, system prompts, and request payloads

Perfect for:

• Building conversational AI interfaces in Mendix apps
• Creating intelligent assistants with streaming responses
• Implementing custom AI workflows that require advanced OpenAI parameters
• Developers who need flexibility beyond standard Mendix AI connectors

This widget solves the complexity of handling streaming responses and advanced API parameters in Mendix, providing a seamless bridge between your low-code platform and cutting-edge AI capabilities.

Documentation

Typical usage scenario

Add a chat interface to your Mendix application that connects directly to OpenAI's API with real-time streaming responses. The widget handles the complexity of Server-Sent Events and provides a ready-to-use chat UI.

Features

✅ Complete chat interface with message history

✅ Real-time streaming responses

✅ Support for Chat Completions and Responses APIs

✅ Configurable through JSON for API parameters

✅ Debug mode for troubleshooting

✅ Trigger Mendix actions on message events

Limitations

❌ Requires OpenAI API key (not included)

❌ No built-in conversation persistence (session-based only)

❌ No file upload/attachment support

❌ Single conversation thread per widget instance

Dependencies

• Mendix Studio Pro 10.x or higher
• OpenAI API key
• Internet connection

Installation

• Import the .mpk file into your project
• Find "OpenAI Chat" in the widget toolbox
• Drag onto your page
• Configure the required properties

Configuration

Required settings:

• API Key: Your OpenAI API key
• API Type: Choose "Chat Completions" or "Responses"

Optional settings:

• Request Configuration: JSON string with model and parameters (default: {"model": "gpt-4", "stream": true})
• System Prompt: Set the assistant's behavior
• Max Messages: Limit displayed messages (default: 50)
• Max/Min Height: Control widget size (e.g., "500px")
• Debug Mode: Log API calls to console

Example request configuration:

'{  "model": "o4-mini",  "reasoning": {    "effort": "high"  }}' 

For more information about OpenAI's API visit the docs: https://platform.openai.com/docs/