Form schema

Open the form's Settings, Fields and schema. If the form already has submissions, click Start from those and WebForms suggests the fields and types it has seen. For each field set a type (text, textarea, email, tel, url, number, date, select, radio, checkbox), whether it is required, the options for a select, and a sentence saying what it is for. That sentence is what AI agents read.

Tick Reject submissions that don't match and a submission with a missing required field, an invalid email or an unknown option gets a 422 listing each problem, before anything is stored or sent. It is off by default, so saving a schema never breaks a live form. Keep fields that aren't listed here decides whether extra fields are kept or refused.

Below the editor: an agent-ready HTML form, TypeScript types, a Zod schema and JSON Schema, each with a copy button. The Zod schema runs the same checks WebForms runs on arrival, so validating in your own code first catches the same mistakes.

export type ContactSubmission = {
  name: string;
  /** Where we reply */
  email: string;
  plan?: "Starter" | "Team";
};

Every form also has a public schema.json and openapi.json at its endpoint, e.g. https://send.webforms.to/your-key/openapi.json, always current, for tools and agents that read API descriptions.

GET /v1/forms/:id/schema returns the fields and every generated output; PUT replaces the fields. Over MCP, an assistant can use get_form_schema and save_form_schema.

curl -X PUT https://api.webforms.to/v1/forms/frm_xxx/schema \
  -H "Authorization: Bearer wf_live_your_key" \
  -H "Content-Type: application/json" \
  -d '{"fields": [
        {"name": "name", "type": "text", "required": true},
        {"name": "email", "type": "email", "required": true, "description": "Where we reply"},
        {"name": "plan", "type": "select", "options": ["Starter", "Team"]}
      ], "enforceSchema": true}'