How It Works

Ev3ry follows a four-step process to go from a website URL to structured, validated data.

Step 1: Add a website

Create a website entry with the URL you want to extract from. Add a short description of what data you're after — this guides the agent's exploration. If the site requires authentication, attach a saved login.

Step 2: Define a schema

A schema is a JSON Schema object that describes the shape of the data you want back.

For example, a match schedule schema:

{
  "type": "array",
  "items": {
    "type": "object",
    "properties": {
      "home_team":  { "type": "string" },
      "away_team":  { "type": "string" },
      "kickoff":    { "type": "string", "format": "date-time" },
      "venue":      { "type": "string" },
      "odds_home":  { "type": "number" }
    },
    "required": ["home_team", "away_team", "kickoff"]
  }
}

Schemas are defined once and can be reused across multiple websites. You can write them manually or paste in a template from the Schemas library.

Step 3: Run the extraction

Click Run on your website page. The AI agent takes over:

1. Opens a browser session (local or cloud)
2. Navigates to your website
3. Explores the page — DOM, network requests, framework globals — to find the best data source
4. Writes an extraction script and validates output against your schema
5. Returns the structured data

You can watch the agent work in real time through the live view — see the browser, the agent's decisions, and data appearing as it's collected.

Step 4: Save a playbook

After a successful run, save the result as a playbook. This captures:

• The navigation steps the agent took
• The extraction script it wrote
• The schema it mapped to

Future runs using this playbook skip the AI exploration phase and go straight to extraction — deterministic, fast, and low-cost.

What comes after

• Download results as JSON or CSV
• Schedule recurring runs via cron expression or webhook trigger
• Chain extractions into a workflow when you need data from multiple websites in sequence