unformal/Docs

Concepts

The building blocks of Unformal — what everything is, how it fits together, and why it works this way.

What is a Pulse?

A Pulse is an AI-powered conversation. Instead of static fields a respondent fills in alone, a Pulse talks with them, adapting questions based on what they say, following interesting threads, and asking the right follow-ups.

You define the intention — what you want to learn — and the AI handles the rest. No flowcharts, no branching logic, no conditional rules to maintain.

Example intention

"Understand what a new client needs before our first meeting — their budget, timeline, current challenges, and what success looks like for them."

Each Pulse gets a shareable link. Share it via email, embed it on a page, or send it from your app via the API. Anyone who opens it starts a Conversation.

What is a Conversation?

Each time someone responds to your Pulse, it creates a Conversation — a back-and-forth dialogue between the respondent and the AI. Conversations collect rich, contextual information that static forms never could.

Respondents can type their answers or use voice dictation — Unformal transcribes speech in any language via Whisper. Most people find talking faster and more natural than typing for open-ended questions.

A conversation can be in one of these states

In progressRespondent is actively answering
CompletedAll questions answered, Echo generated
AbandonedInactive for 24+ hours without completing

What is an Echo?

When a conversation completes, the AI analyzes the entire transcript and produces an Echo — a structured summary that distills everything said into actionable output.

The Echo is what you actually work with. It contains:

  • —Structured fields — The specific data points you asked for — extracted and typed (see Output Fields below)
  • —Summary — A concise narrative of the conversation
  • —Key quotes — The most important things the respondent said, verbatim
  • —Subtext — What the AI inferred beyond the literal answers — hesitations, enthusiasm, unstated concerns
  • —Sentiment score — A 1–10 rating of how the respondent felt about the topic
{
  "echo": {
    "fields": {
      "budget_range": "$50k–100k",
      "timeline": "Q3 2026",
      "pain_points": ["manual data entry", "no single source of truth"]
    },
    "summary": "Strong fit. Budget and timeline aligned.",
    "keyQuotes": ["We spend 3 hours daily on data entry"],
    "subtext": "Enthusiastic but worried about internal buy-in.",
    "sentimentScore": 7
  }
}

Interview Mode vs Extract Mode

Unformal supports two conversation modes. Same output structure, different experience for the respondent.

🎙️

Interview Mode

default

The AI leads. It asks one question at a time, listens to each answer, and decides what to ask next. It follows interesting threads, asks natural follow-ups, and explores topics in depth.

Best for:

Client intakeUser researchFeedback collectionQualitative discovery
🗒️

Extract Mode

The respondent leads. They write everything they want to say upfront — no structure required. The AI then reads what they wrote, identifies what's missing from your output fields, and asks only targeted follow-up questions to fill the gaps.

Faster and less friction for people who already know what they want to say. The AI does the work of structuring it.

Best for:

Brain dumpsBug reportsQuick intake from verbose inputTechnical stakeholders

Tones

The tone controls how the AI talks to your respondent. Pick one that fits your audience and context. You can override it any time in the Pulse settings.

🎯

Professional

Formal and structured. Works well for enterprise client intake, compliance, and any context where credibility matters.

🌱

Coaching

Supportive and probing. Encourages reflection, follows up with empathy. Great for 1:1 feedback and wellbeing contexts.

💬

Conversational

Natural and relaxed. Feels like talking to a knowledgeable colleague. The default for most use cases.

😄

Casual

Friendly and informal with a touch of humor. Reduces friction for respondents who'd rather not fill in forms.

🤖

Self-aware

Slightly meta and witty. The AI knows it's an AI and doesn't pretend otherwise. Good for technical audiences.

Extract mode defaults to Casual — a brain dump feels less formal when the AI matches that energy.

Output Fields (Echo Schema)

Output fields define the structured data the AI should extract from each conversation. They're the bridge between a free-flowing dialogue and the clean, typed data you actually need.

Each field has:

name

snake_case identifier

"budget_range"

type

Data type

"string" | "number" | "boolean" | "array"

description

What to extract

"Estimated annual budget range"

The AI reads this to know what to look for

// Example output fields for a client intake Pulse

"outputFields": [
  {
    "name": "budget_range",
    "type": "string",
    "description": "Estimated annual budget range they can work with"
  },
  {
    "name": "decision_maker",
    "type": "boolean",
    "description": "Whether this person can make the buying decision alone"
  },
  {
    "name": "pain_points",
    "type": "array",
    "description": "List of their top 3 challenges or frustrations"
  },
  {
    "name": "timeline_weeks",
    "type": "number",
    "description": "How many weeks until they need a solution"
  }
]

Fields are auto-generated from your intention when you click "Generate" in the Pulse creator. You can adjust them before publishing.

Resonance

Once a Pulse has 3 or more completed conversations, Unformal generates Resonance — aggregate intelligence across all responses.

A single conversation tells you what one person thinks. Resonance tells you what everyone thinks — and where they agree, disagree, or feel differently.

—

Common themes

Topics and concerns that appear repeatedly across conversations

—

Consensus points

Things most respondents agree on

—

Divergent opinions

Areas where respondents meaningfully disagree

—

Sentiment distribution

Breakdown of how positive/neutral/negative the responses are overall

—

Recommended actions

AI-generated suggestions based on the aggregate patterns

Resonance is visible on the Studio Resonance tab. It updates automatically as new conversations complete.

Pricing

Every completed conversation counts as one. In-progress or abandoned conversations are free. Research-enabled conversations count as two.

Every new account starts with 50 free credits. Then add a one-time pack: €9 for 100, €29 for 500, €79 for 2,000 or €249 for 10,000 credits.

€0

50 credits to start

Free

€9

100 credits, one time

Starter

€29

500 credits, one time

Growth

€79

2,000 credits, one time

Scale

Every pack unlocks Pro permanently, including watermark removal. Credits never expire. Standard completed conversations use 1 credit, or 2 with research.