Migration Assistant for API Responses

62 views 0

Introduction

Notice of end of support for the Assistants API.

The Responses API is OpenAI's most modern interface for building AI applications and is gradually replacing the Assistants API . It offers a simpler, more flexible, and agent-ready integration model, unifying text generation, multimodality, tool calls (Web Search, File Search, Computer Use, and Function Calling), and context management into a single API.

In comparison, the Assistants API was designed around the concepts of Assistants , Threads , and Runs , abstracting conversation management. While it still works for existing integrations, it has been officially discontinued for new projects and is scheduled to be shut down on August 26 , 2026. OpenAI recommends that all new applications use the Responses API.

  • Responses API: Simpler architecture, greater flexibility, support for the latest features, and a focus on the platform's future.
  • Assistants API: A model based on persistent assistants and threads, suitable for legacy applications, but currently being phased out.

Migration of Assistants in ByteGPT

When opening the widget in KommoInstructions will then appear for migrating to the new OpenAI API model.

Step 1: Get instructions from the assistant.

  1. Access the assistant that you imported from the list of providers. ByteGPT.
  2. Click on the button "Seek instructions from the Assistant" so that the ByteGPT Search for available AI models and instructions from your Assistant.

Step 2: Review the instructions and select the AI ​​model.

Migration to API responses
  1. Select an AI model for your assistant.
  2. Review the instructions provided by OpenAI. ByteGPT.
  3. Activate Vector Stores (optional). 

Assistant files

The files will not be imported into the ByteGPTIf you use them, you will need to create a Vector Store In OpenAI, attach your files.

When creating a Vector Store and adding files in OpenAI, click the button again. "Seek instructions from the Assistant" so that they appear in ByteGPT.

Instructions for adding files to OpenAI vector files.

Step 3: Save and migrate the Assistant

After reviewing the template, instructions, and files, simply click the "Save and Migrate Assistant" button.

After completing the migration, you will be able to add new files to OpenAI and synchronize them. ByteGPT clicking the button "Synchronizing OpenAI Vector Stores".

The instruction adjustment can now be made directly in ByteGPTWhen finished, click on "Save Settings".

Assistant migrated to the Responses API.

Add an AI from the new API model.

If you are not using the OpenAI Wizards and want to perform the configuration from scratch, follow these steps:

Step 1: Registering the Responses API in ByteGPT

Selecting Responses API in ByteGPT
  1. Please click  "Add" in AI Providers.
  2. Select OpenAI (Responses).
API key and assistant name in the settings of ByteGPT

3. Enter the OpenAI API Key .

4. In Name/Identifier, enter a name to identify your AI. Example: AI of ByteGPT.

Step 2: Configure and synchronize instructions and files.

  1. Please click  "Configure/Synchronize" protocols for ByteGPT Retrieve the AI ​​models from your OpenAI account.
Migration to API responses

2. Change the AI ​​model to the desired one.

3. Add the instructions.

4. Select a file from Vector Stores (optional).

5. Click on “Save Settings”.

Step 3: Import the template from ByteGPT

Templates ByteGPT
  1. Download the template from ByteGPT Which one best suits your needs? See the differences between each one. in this guide. 
  2. Import the downloaded template into Kommo.
After downloading the template, go to “Settings > Communication tools > Salesbots”, click on “import a new bot” and import the downloaded template.

What this helpful?

 
Related articles