Loading vibe-surveyPreparing the next workspace surface.
Concept Test demo · HTML DSL · GUI · runtime contracts
Survey programming, readable as code.
Build research surveys from inspectable HTML source, edit them in GUI, and keep the structure clear enough for AI agents to inspect.
Primary proof
Source, editor, and runtime stay on one board.
Concept Test demoQuestion codes aligned
survey.htmlHTML DSL
1<question type="single" name="SQ1" required>2 <description>Consent</description>3 <options order="asc">4 <option value="1">Continue</option>Runtime attachment
Question codes stay alignedNames, options, and display-style remain visible.
Runtime stays attachedLogic, quota, Survey Link, and export keep the same map.
Editor
GUI syncQ6_{{item}} · Single choice
Appeal for {{concept_title}}
Very appealing1
Somewhat appealing2
Neutral3
Loop, quota, and Survey Link continue from the same question names and option values.
Editor
GUIHTMLStyle
SQ1Single · Consent
SQ2Number · Age
Q3Hidden multi
Q8Ranking · accordion
Q6_{{item}} · Single choice
Appeal for {{concept_title}}
Very appealing1
Somewhat appealing2
Neutral3
Not appealing4
Loop: CT_CONCEPT_LOOP · source: hidden Q3 concept assignment
Loop
Named items stay explicit.
Quota
Workbook cells remain attached.
Survey Link
Runtime renders the same map.
Quota · Run generate · Survey Link
Mobile proof keeps the contract visible without stacking every desktop surface.
Product philosophy
The survey source is the contract.
Author in readable HTML, refine in GUI, and keep every runtime surface attached to the same structure.
Readable source
Survey structure lives in HTML that researchers and engineers can inspect.
Aligned editor
GUI changes stay attached to the same names, option values, and display styles.
Operational proof
Quota, response, Survey Link, and export behavior remain part of the contract.
AI-native by structure
Readable surveys make useful agents possible.
The agent does not guess from a screenshot. It reads names, codes, conditions, loops, quotas, and runtime contracts.
Concept Test runtime mapHTML DSL
1<survey id="concept_test">2<quota source="concept_quota" title="Concept assignment priority quota" />3<loop name="CT_CONCEPT_LOOP">4 <question type="ranking" name="Q8_{{item}}" display-style="accordion" />5 <question type="number" name="Q10_{{item}}" display-style="slider" />6 <question type="nps" name="Q11_{{item}}" display-style="carousel" />7</loop>8</survey>Question Object
Quota DSL
Runtime
Export shape
Product proof
Every operating surface stays legible.
Source
Inspectable
Editor
Aligned
Runtime
Named
Export
Predictable
runtime-map.htmlHTML DSL
1<survey id="concept_test">2<quota source="concept_quota" title="Concept assignment priority quota" />3<loop name="CT_CONCEPT_LOOP">4 <question type="ranking" name="Q8_{{item}}" display-style="accordion" />5 <question type="number" name="Q10_{{item}}" display-style="slider" />6 <question type="nps" name="Q11_{{item}}" display-style="carousel" />7</loop>8</survey>Workbench
Test + LiveBlocks
ConsentConcept loopQuota helper
Selected question
Q10_{{item}} · Slider number
Mode: float · Display-style: slider · Validation: 0–100
Live diff token→Quota workbook→Survey Link→Export names
The same identifiers move from authoring through respondent runtime and export without becoming image-only proof.
HTML DSL
Names, option codes, display-style, and authored logic.
GUI
Editor controls for blocks, question types, options, and settings.
Logic
Screen-outs, conditions, loops, and run-code stay inspectable.
Quota
Assignment cells and quota helpers connect to runtime behavior.
Run generate
Simulation checks the survey map before live collection.
Survey Link
Respondent pages render from the same source contract.
Responses
Collected answers retain source, quota, and version context.
Export
Response names, versions, and metadata stay predictable.
Core surfaces
One source. Every operating surface nearby.
Author, validate, preview, quota, collect, review, and export without splitting the survey into disconnected tools.
Runtime chain
LogicConditions, loops, and screen-outs remain named and testable.
QuotaAssignment cells stay linked to the live contract.
Run generateSimulation checks surface drift before collection.
Survey LinkRespondent rendering reads the same survey map.
ExportResponse names and metadata stay predictable.
Prefer crisp authored proof over schematic placeholders.
Keep source, GUI, runtime, and export legible on desktop and mobile.
Use circuit motion only as ambient guidance, never as the main story.
AI-readable source boundaryHTML DSL
1<survey id="concept_test">2<quota source="concept_quota" title="Concept assignment priority quota" />3<loop name="CT_CONCEPT_LOOP">4 <question type="ranking" name="Q8_{{item}}" display-style="accordion" />5 <question type="number" name="Q10_{{item}}" display-style="slider" />6 <question type="nps" name="Q11_{{item}}" display-style="carousel" />7</loop>8</survey>Condition graph
StartSQ1Screen-out
Export shape
responses · quota_table · version · metadata
Agent interfaces
Give agents structure they can safely inspect.
Skills, MCP, and app surfaces can act on survey maps only because the source remains explicit.
Codex Skill
Programmatic access for reasoning over survey structure and rules.
Claude Skill
Inspect, summarize, and author surveys from explicit source.
MCP
Model Context Protocol endpoint for survey maps and operations.
ChatGPT Apps
Embedded app surface for building surveys inside agent workflows.
vibe-surveyHTML DSL survey builderStart with the demo. Inspect the source when logic matters.
HTML DSL, GUI editing, quota operations, Survey Link, responses, and export stay grounded in one source contract.
Source statusv1.0.0Concept Test demo Survey Link ready