Survey Creator

Lay out questions as blocks, reorder them, preview the respondent view, then export JSON or keep a draft in local storage. Built for fast outlines before you move work into email, slides, or a dedicated form product.

0 blocks~0 min read

No blocks yet. Tap a chip below to insert your first field. On small screens the dock at the bottom mirrors preview, export, and save.

Pick blocks before you polish copy

Most bad surveys fail in the outline stage, not the font choice.

Start by dropping one block per intent: a short text field for a name, a one-choice row for a priority bucket, a Likert strip for agreement, a date for a deadline. You reorder with the chevrons on each card so the story reads top to bottom the way a tired respondent would skim it on a phone. Once the skeleton feels honest, rewrite prompts so each question earns its place.

We recommend exporting JSON when you need to hand the structure to a teammate or archive a version. The file is plain text; you diff it in git or paste keys into another system. If you only need a scratchpad, Save draft writes into local storage for this browser profile.

Nothing here syncs to a server. Preview, export, and saved drafts stay on the machine where you run the page. Clear site data and drafts disappear, so treat exports as your durable copy.

Which block for which job

Use the table as a cheat sheet when you hesitate between two shapes.

BlockBest forWatch out
Short textNames, IDs, one-line answersOpen text is hard to chart later
ParagraphFeedback stories, bug reportsLong answers need manual reading
One choiceExclusive categoriesOverlapping options confuse counts
Multi selectFeature picks, symptom listsToo many boxes fatigues people
DropdownLong option listsMobile users dislike tiny menus
Number scaleQuick satisfaction scoresDefine what 1 and 5 mean
Likert rowAgreement without long proseKeep labels parallel in grammar
DateDeadlines, availabilityTime zones still need prose

Friction you should plan for

Even a clean outline bumps into real-world mess.

If you already captured messy CSV exports from another tool, normalize headers and blank rows before you build charts. The Data Validator catches structural problems early so you do not argue about counts in a meeting.

What stays in the tab (and what does not)

This page does not collect responses from the public, send email, or enforce skip logic. You get structure only.

Branching rules, quotas, and identity checks belong in a dedicated collector. Treat the creator here as a whiteboard. When you need a lighter vote-style flow without the full questionnaire shape, the Poll Generator is often the faster path. When you already have CSV or spreadsheet answers and need consistency checks before analysis, run them through the Data Validator after export.

Teams debating wording before they freeze a form sometimes pair this outline with the Decision Maker for quick tie-breaks on question order or scale width.

Stop when the outline is clear

Twenty blocks rarely beat eight sharp ones.

Ship the JSON or draft, test with three colleagues who resemble real respondents, then cut anything they misread. The preview button is deliberately read-only so you notice awkward phrasing instead of chasing fake submit metrics inside the builder.

Survey Creator: quick answers

Practical notes about drafts, files, and limits.

Where do saved drafts live?

Save draft stores JSON in your browser local storage for this site. It does not sync across devices or browsers. Export JSON if you need a file you can move or back up.

Does preview collect answers?

No. Preview renders a static respondent view with disabled controls so you can read the flow. Nothing posts to a server from this page.

Can I import JSON back into the builder?

Not today. Import is a logical next step, but the current tool only builds forward. Keep exported files as your source of truth until an import path exists.

How do I share the survey with people?

Export the structure and recreate it in whatever system you use to host forms, or paste questions into email or docs if the audience is tiny. This page is not a hosted collector.

Why do block counts affect the time pill?

The estimate assumes roughly half a minute per block for an average respondent skimming on mobile. It is a planning hint, not a statistical model.