QubeSurvey
Write and refine questionnaires with your chosen AI agent. QubeScript keeps questions, answer options and routing explicit; the local MCP supplies language guidance and checks.
Test locally with the QubeSurvey extension, review with your team, then run fieldwork and export the responses.
Use your preferred AI agent for authoring, with testing, fieldwork and data together in QubeSurvey.
Download and install the QubeSurvey VS Code extension using the extension guide. Use its local MCP with an AI agent you choose to draft, explain, check and edit QubeScript. Test interview paths locally with invented answers.
Upload a reviewed script for team testing in the browser. File tickets against that version with the expected behavior and available interview history. Review and test proposed fixes before using them.
Create a Run from a reviewed version, review its settings, activate it and share its survey link.
Review Run responses and create an available CSV, JSONL or SSS export. Check the exported rows and data dictionary.
Specification and source
Questions, answer options, routing and validation live in one readable script. You and your AI agent can inspect the questionnaire and propose changes with that context.
The local MCP supplies language guidance and checks. Test the interview in the extension before sharing it with your team. QubeSurvey also supports Odin: run Odin surveys on QubeSurvey, or audit and test a script before running it on NIPO Fieldwork System.
@question q_purchase_p4w multi
Which of these brands have you purchased in the **past 4 weeks**?
@options grain_brands.select(answers.q_aided_awareness.ids(), 98, 99)
@route purchase_branch
if answers.q_purchase_p4w.contains(99) -> consideration_section
else -> most_often_section
Team testing
Upload a version for your team to test in the browser. Tickets keep the tested version, expected behavior and recorded interview history when available. Odin and QubeScript browser tests support saved test journeys.
The hosted Assistant can propose a corrected version. Compare the source, then test the affected and nearby routes before using it.
Hosted fieldwork
A Script holds the questionnaire and its versions. A Run selects one version and controls entry, collection dates and which responses belong together.
Create a Run from a reviewed version, review its settings, activate it and share its survey link. Run responses can be exported as CSV, JSONL or SSS.
Pricing
Writing scripts, local interview testing, team testing in Script preview, and filing manual tickets are free. One prepaid Account balance covers Run responses, survey email and hosted AI.
Free
Your chosen editor agent’s provider may charge separately.
Extension installation$0.10 USD / accepted response
Review AI spending limits before starting a task. Sign-in and billing notices are free and rate-limited.
View billingGet started
Start with an existing study specification, a routing requirement or a new research question. Write and test locally before uploading a version.
Start with a study specification and its existing instrument. Test Odin routes locally, run supported Odin surveys on QubeSurvey, or audit and test a script before NIPO Fieldwork System.
Audit or run an Odin script →Work from an example with screening, follow-up choices, loops or calculations. Adapt it to your study, test different answers and check what happens when an earlier answer changes.
Explore survey logic →Name one decision and target group. Ask your assistant to draft questions, test the flow with invented answers, then review the wording, audience and recruitment plan yourself.
Draft your first survey →