client.questions

In linear mode, the interviewer asks these questions in order, probing with follow-ups. For branching interviews, see client.graph.

list()

client.questions.list(study_id) -> QuestionListOut

The study’s questions, in interview order. Returns QuestionListOut.

create()

client.questions.create(study_id, *, question: str, qtype: str | None = None, **config) -> QuestionOut

Create a question and append it to the end of the list.

str | UUID

required

The study.

str

required

Question text.

str

text, scale, choices, multi_select, ranking, slots, card_sort or image_upload. If omitted, it’s inferred: scale → scale, options → choices, slots → slots, expected_image → image_upload, otherwise text. image_upload needs the standard or premium model tier.

str

Guidance for the interviewer on how to probe this question.

ScaleConfig | dict

For scale questions, for example {"minScale": 1, "maxScale": 5, "minLabel": "Not at all", "maxLabel": "Very"}.

list[str]

Answer options, for choices, multi_select and ranking.

list[str]

Named slots to fill, for slots questions.

list[str]

Bucket names, for card_sort questions.

int

Minimum selections, for multi_select.

int

Maximum selections, for multi_select.

bool

default:"True"

Whether the interviewer asks follow-up questions.

int

Minimum conversation turns on this question.

int

Maximum conversation turns on this question.

str

Description of the expected upload, for image_upload questions.

Returns QuestionOut.

from deutero.models import ScaleConfig

client.questions.create(
    study.id,
    question="How easy was setup?",
    scale=ScaleConfig(minScale=1, maxScale=5, minLabel="Very hard", maxLabel="Very easy"),
)

client.questions.create(
    study.id,
    question="Rank these by how much they matter to you.",
    qtype="ranking",
    options=["Price", "Speed", "Support", "Integrations"],
)

client.questions.create(
    study.id,
    question="Sort these features by how often you use them.",
    qtype="card_sort",
    options=["Reports", "Alerts", "API", "Exports"],
    groups=["Daily", "Weekly", "Rarely"],
)

client.questions.create(
    study.id,
    question="Share a screenshot of your current dashboard.",
    expected_image="A screenshot of an analytics dashboard",
)

generate()

client.questions.generate(
    study_id, *, n_questions: int | None = None, additional_instructions: str | None = None
) -> GeneratedQuestionsOut

Write questions with AI and append them to the list. The generator works from the study’s research question, objectives, methodology and target population, picks a mix of question types and writes in the study’s language. New questions extend the existing list; nothing is replaced. Usually takes 30 seconds to 2 minutes.

int

default:"10"

How many questions to generate, from 3 to 20.

str

Extra guidance, such as "start with two warm-up questions".

Returns GeneratedQuestionsOut. skipped_image_questions counts image-upload questions the generator proposed but didn’t add, because your plan has no vision-capable interview model.

update()

client.questions.update(study_id, question_id, **fields) -> QuestionOut

Partially update a question. Takes the same optional fields as create(), including question; only the ones you pass change. Pass an empty list to clear options, slots or groups. Returns QuestionOut.

delete()

client.questions.delete(study_id, question_id) -> SuccessResponse

Delete a question and its attached images. The remaining questions are renumbered. Returns SuccessResponse.

reorder()

client.questions.reorder(study_id, *, question_ids: Sequence[str | UUID]) -> QuestionListOut

Pass every question ID, in the interview order you want.

questions = client.questions.list(study.id).questions
client.questions.reorder(study.id, question_ids=[q.id for q in reversed(questions)])

Returns QuestionListOut.

validate()

client.questions.validate(study_id) -> ValidationOut

Review the questions before going live: an overall ethics pass or fail, plus per-question language quality and redundancy findings. It’s the same check as the dashboard’s Validate button. For a check of the whole study before publishing, use studies.validate().

report = client.questions.validate(study.id)
print("Ethics passed:", report.ethics_check.passed)
for q in report.questions:
    if q.has_issues:
        print(q.number, q.language_issues, q.redundancy, q.suggestions)

Returns ValidationOut.

Client

Study setup

Recruitment

Simulation

Results

Organization

Utilities

list()

create()

generate()

update()

delete()

reorder()

validate()

​list()

​create()

​generate()

​update()

​delete()

​reorder()

​validate()

list()

create()

generate()

update()

delete()

reorder()

validate()