Your forms, supercharged with voice.
Fill layers onto your existing web forms — no rebuild required. Users speak naturally, and Fill maps their words to the fields you already have. One script tag. Zero changes to your UI.
From voice to structured data in three steps.
Users click the mic and describe what they need in natural language.
Audio is transcribed and mapped to your form’s JSON schema by an LLM.
Each field populates automatically with a per-field confidence score.
Everything you need for voice-powered forms.
Voice-to-Form
Users speak naturally and Fill maps their words to your existing form fields. No retraining, no new UI.
JSON Schema Mapping
Define your form structure once as JSON schema. Fill understands field types, enums, and validation rules.
Confidence Scoring
Every extracted field gets a confidence score. Flag low-confidence values for human review before submission.
Multi-Provider STT
Choose your transcription engine: OpenAI Whisper, Deepgram, or Google Speech-to-Text. Swap anytime.
Bring Your Own Keys
Connect your own LLM and transcription API keys. You control the model, the cost, and the data.
Zero UI Changes
Fill is a lightweight overlay that attaches to your existing forms. One script tag, no rebuild required.
Built for forms that slow people down.
Policyholders describe incidents verbally and the claim form populates instantly.
Front desk staff speak patient details instead of typing through long intake forms.
Employees dictate receipts and Fill extracts vendor, amount, date, and category.
New customers speak through profile setup instead of tabbing through form fields.
<script src="kind-robots-fill.js" data-form-id="claims_intake"></script>