Introduction

Building an AI chatbot has never been simpler — or more powerful. With the release of OpenAI’s Responses API a few months ago, developers can now create stateful, multimodal chat experiences using a single, unified interface. In this article, we’ll walk through a simple AI chatbot built with Next.js that leverages this new API to handle both text and image generation, while maintaining conversation state effortlessly.

Why the Responses API Changes Everything

The Responses API is OpenAI’s most advanced endpoint for generating model outputs. It’s designed to replace older APIs, such as the Chat Completions API and the Assistants API, streamlining the developer experience with features like:

• Text and image inputs
• Text outputs
• Built-in tools such as image generation, file search, web search, and function calling
• Easy conversation state management
• Support for external systems via MCP servers

This means you can now build rich, interactive AI applications without juggling multiple endpoints or manually managing context.

Key Features Used in Our Chatbot

Conversation State

One of the standout features of the Responses API is its ability to track conversation state using the previous_response_id parameter. Instead of passing the entire message history, you simply reference the last response ID, and the model retains context across turns.

This makes it incredibly easy to build threaded conversations without bloating your payloads.

Image Generation with Built-in Tools

The chatbot also uses the image_generation tool — natively supported by the Responses API. No need to call a separate image endpoint. Just include the tool in your request, and the API can generate images directly.

Final Thoughts

The OpenAI Responses API marks a major leap forward in building intelligent, multimodal applications powered by AI. With built-in tools, easy context management, and unified input/output handling, it’s now easier than ever to create extendable and customizable chatbots and AI agents — and it will probably get even better in the future.

Resources

The AI Chatbot source code is available on GitHub.

To dive deeper into the Responses API, check out OpenAI’s official documentation.