Is Sequential Function Calling Possible in the Responses API Like in the Assistants API?

API
1

Hi everyone,

I wanted to ask if it’s possible to call multiple functions sequentially using the Responses API, similar to how function/tool chaining works in the Assistants API.

Specifically, I’m looking to implement a flow where:

  1. One function is called first based on the user’s input,
  2. Its output is then passed as input to a second function,
  3. And this continues for several steps in one interaction or conversation cycle.

I’m aware that the Assistants API supports tool chaining where the assistant can automatically call tools in sequence based on the plan. But I’m not sure if the Responses API supports something similar — or if this would need to be managed entirely on the application side.

Would love any guidance, best practices, or even example implementations from others who’ve tackled this.

Thanks in advance!

Wasif

2

Hi, did you figure it out?

3

If it is a mandatory path that doesn’t need intermediate AI thinking, you can just chain the tools yourself with the full inputs - or better, have the tool be AI powered and completely internally prompted.

Otherwise, you can add prohibitions to the dependent function description: that it cannot be invoked without the results of the first tool, and cannot be sent in parallel.