Hi all,
I'm trying to leverage 2 parameters of the OpenAI Responses API: prompt and previous_response_id.
I have the code below that works but unfortunately then the output of all my tests is just the response ID instead of being the OpenAI text response.
Is there an option to save the response ID while still having the test return the text?
Here's the relevant parts of our setup:
```
prompts:
- "dummy"
providers:
- id: https
config:
url:
https://api.openai.com/v1/responses
method: POST
headers:
Authorization: "Bearer {{ env.OPENAI_API_KEY }}"
body: |
{
"model": "gpt-4.1-nano-2025-04-14",
"input": "{{ tenant_msg }}",
"prompt": {
"id": "{{ promptId }}"
},
{% if previous_response_id %}
"previous_response_id": "{{ previous_response_id }}",
{% endif %}
"store": true
}
transformResponse: |
({
output: {
text: json.output[0].content[0].text,
response_id: json.id
}
})
defaultTest:
vars:
previous_response_id: ""
tests:
- description: "1 • tenant contacts"
vars: {
tenant_msg: "Good morning.",
promptId: "pmpt_685a89e16a208196bb088d0b09cbee3f09404e8cfa6d680c"
}
options:
transform: output.response_id
storeOutputAs: previous_response_id
- description: "2 • tenant follow-up"
vars: {
tenant_msg: "Did you get my previous message?",
promptId: "pmpt_685a89e16a208196bb088d0b09cbee3f09404e8cfa6d680c"
}
options:
transform: output.response_id
storeOutputAs: previous_response_id