how to use Code Interpreter or Image Output in OpenAI Agents SDK

[tasif498]

Is there currently a way to use the Code Interpreter (Python execution) tool or return images (e.g. charts) in responses when using the OpenAI Agents SDK?
If not, I’d like to request one of the following:
Option 1 Code Interpreter Integration
Support for securely executing Python code (like in the Assistants API with the code interpreter tool) to enable dynamic data processing and chart generation.

Option 2 Image Output Support
If code execution isn't feasible, provide a way for agents to return image outputs (e.g. base64-encoded charts) in their responses. This would allow external tools to handle code, and the agent simply returns the result.

Use Case:
Building data analysis agents that can respond to user queries with summaries and visual charts (bar, pie, line, etc.).

[pierluigi-D-segatto]

is there anyone that can answer on how/when these features can/will be implemented?

[github-actions]

This issue is stale because it has been open for 7 days with no activity.

[tasif498]

Can we get any update on this?

[bruno-curta]

I'm also struggling with this point.
I have tried to create assistants and tools to use code interpreter to generate the HTML based on the encoded image URL, but it is not working. This should be easier to configure, I believe.
Any updates on that?

[portizv26]

I am struggling with this as well.
My initial approach is to orchestrate a flow in which an assistant creates an image an uploads it into a repository that can be later be reached by my agent, but I don't know if this approach is secure or if there is a better approach for this.

[github-actions]

This issue is stale because it has been open for 7 days with no activity.

[rm-openai]

Hey I'm so sorry I totally missed this. I don't think we have a code-interpreter hosted tool. I can add some image examples. Can you share some sample code and comment on parts where you're not sure what to do?

[isaac47]

Hi, Can you please give an example of code showing agent taking image as input to generate description ?

[rm-openai]

@isaac47 #553

[github-actions]

This issue is stale because it has been open for 7 days with no activity.

[github-actions]

This issue was closed because it has been inactive for 3 days since being marked as stale.