Today, we’re releasing the Assistants API, our first step towards helping developers construct agent-like experiences inside their very own applications. An assistant is a purpose-built AI that has specific instructions, leverages extra knowledge, and may call models and tools to perform tasks. The brand new Assistants API provides recent capabilities similar to Code Interpreter and Retrieval in addition to function calling to handle a whole lot of the heavy lifting that you just previously needed to do yourself and enable you to construct high-quality AI apps.

This API is designed for flexibility; use cases range from a natural language-based data evaluation app, a coding assistant, an AI-powered vacation planner, a voice-controlled DJ, a wise visual canvas—the list goes on. The Assistants API is built on the identical capabilities that enable our recent GPTs product: custom instructions and tools similar to Code interpreter, Retrieval, and performance calling.

A key change introduced by this API is persistent and infinitely long threads, which permit developers at hand off thread state management to OpenAI and work around context window constraints. With the Assistants API, you just add each recent message to an existing thread.

Assistants even have access to call recent tools as needed, including:

• Code Interpreter: writes and runs Python code in a sandboxed execution environment, and may generate graphs and charts, and process files with diverse data and formatting. It allows your assistants to run code iteratively to resolve difficult code and math problems, and more.
• Retrieval: augments the assistant with knowledge from outside our models, similar to proprietary domain data, product information or documents provided by your users. This implies you don’t must compute and store embeddings to your documents, or implement chunking and search algorithms. The Assistants API optimizes what retrieval technique to make use of based on our experience constructing knowledge retrieval in ChatGPT.
• Function calling: enables assistants to invoke functions you define and incorporate the function response of their messages.

As with the remaining of the platform, data and files passed to the OpenAI API are never used to coach our models and developers can delete the information after they see fit.

You possibly can try the Assistants API beta without writing any code by heading to the Assistants playground.