Questra
Sign inStart free

Catch survey errors before fieldwork.

Add your survey’s test link and its questionnaire. AI respondents take the survey and show you what passed, what failed, and why, wherever it was programmed.

Word (.docx), text, Markdown, CSV or HTML · up to 5 MB · You’ll add the test link next. Why your questionnaire?

Link test · Example coffee study

AI respondents are working through your survey.

0 of 22 checked
  • S1: “Never” ends the survey. Checking
  • S2: Under-18s are screened out. Waiting for testing
  • S3: Question text matches the questionnaire. Waiting for testing
  • Q1: Every listed brand appears, plus Other. Waiting for testing
  • Q1: “Other” asks the respondent to specify. Waiting for testing
  • Q1: Brand order is randomized. Waiting for testing
  • Q2: Shows only the brands chosen in Q1. Waiting for testing
  • Q2: Carried-forward brands keep their order. Waiting for testing
  • Q3: Pipes the brand from Q2 into the question. Waiting for testing
  • Q4: The scale runs 1 to 10 with both end labels. Waiting for testing
  • Q5: “None of these” stays last. Waiting for testing
  • Q5: “None of these” is exclusive. Waiting for testing
  • Q6: Asked only when Q4 is 7 or higher. Waiting for testing
  • Q7: Answer options match the questionnaire. Waiting for testing
  • Q8: Allows up to three answers. Waiting for testing
  • Q9: Grid rows rotate. Waiting for testing
  • Q10: The open end is required. Waiting for testing
  • MD1: Shows four attributes per task. Waiting for testing
  • MD1: Most and least must differ. Waiting for testing
  • D1: Region list matches the questionnaire. Waiting for testing
  • D2: “Prefer not to say” is anchored last. Waiting for testing
  • END: Completes reach the thank-you page. Waiting for testing

HOW IT WORKS

Your questionnaire is the spec. We test against it.

Questra turns the requirements in your questionnaire into tests, then sends synthetic respondents through your survey to check each one. That’s why every test starts with your questionnaire.

You add the test link MinutesEvery test has a verdict
  1. 01

    Test extraction.

    Questra reads your questionnaire and turns every requirement into a test.

    What it tests
  2. 02

    Scenario planning.

    It plans a respondent path for every answer that changes the route.

  3. 03

    Agent swarm.

    Synthetic respondents take every path on your test link at once.

    Watch an agent
  4. 04

    Test verdicting.

    Each test passes or fails, with the screens the agents saw as evidence.

    See a sample report

WHAT IT TESTS

Four kinds of test, all from your questionnaire.

Every requirement becomes a test in one of four groups, and each one is judged on its own.

  • Routing & eligibility

    Screen-outs, skips, and who sees each question, plus exclusive answers, selection limits, and required inputs.

    • S1 · “Never” ends the survey
    • Q4 · Asked only of Q1 buyers

  • Content & display

    Question and answer wording, piping, randomized and anchored answer order, and emphasis.

    • Q1 · Brand order is randomized
    • Q3 · Pipes the Q2 brand

  • Data quality

    Answer codes, recorded values, recodes, hidden variables, and other-specify text.

    • Q1 · “Other” text is recorded
    • S2 · Ages record as codes
    Needs a connected survey

  • Quotas

    Quota assignments, and what happens when a quota is full.

    • QF1 · A full quota ends the survey
    Needs a connected survey

SYNTHETIC AGENTS

One agent per path, clicking through your survey.

Each agent opens your test link, answers as its path requires, and records every screen it reaches. Pick a respondent to watch its route.

Your survey · test link
PLANNED PATHS
Under 18
Testing…
18–64
Up next
65+
Up next

Selecting an answer. Following the route.

RE-RUNS

Re-run every test up to 5 times per survey.

Fix what failed in your platform, then run the full suite again. Each run shows what your fixes resolved, so you can confirm every round before your first respondent.

Coffee habits2 of 5 runs left
  1. Run 1
    31/50
  2. Run 2
    44/50
  3. Run 3
    50/50
Re-run tests

FIX WITH AI

Fix failed tests with AI in Survey Programming.

With Survey Programming, Fix with AI hands a failed test to the agent, which fixes the survey’s code and checks it again. Link Testing is also half price.

Explore Survey Programming

QUESTIONS

Link testing, explained.

What you need, how long a test takes, and how runs are billed.

Do I need to move my programming to Questra?

No. Start from your questionnaire and your survey’s test link. Your team or programming partner keeps its current platform and tools.

Can I use my live fielding link?

Use a test or QA link. Agents submit answers as they test, so a live link would collect test responses.

How long does a test take?

Minutes. Agents walk the planned paths in parallel, and each check reports as soon as the paths it needs are finished. Surveys with more routes take longer.

Will Questra fix the issues it finds?

Link Testing reports each issue with its evidence, for your team to fix in your platform. When Questra also programs the survey, a QA specialist tests and fixes it before Review, and Fix with AI hands any failed check to the agent, which fixes the source and re-checks it.

What counts as a run?

One full test of your survey against every check. A run counts once its first browser opens, including a run you stop early. A run that cannot start does not count.

How is Link Testing billed?

Link Testing costs your survey’s tier in credits, from your workspace’s shared balance. One purchase covers 5 full runs, and each re-run uses one. When they are used up, buy 5 more. It is half price when Questra also programs the survey. The free test on this page is one run per verified email.

TEST YOUR SURVEY

Go live with confidence in your survey.

Try a complimentary test with your questionnaire and your survey’s test link. One run per verified email.

Try a complimentary test Create a free account

No credit card required.