Parser.run

Understand the building blocks of Parser Run: projects, schemas, jobs, and data.

Projects

A project is a scraping configuration targeting a specific data goal. Each project has its own database, schema, and schedule.

Example: Price Tracker Project

{
  id: "nike-price-tracker",
  name: "Nike Price Tracker",
  description: "Track Nike shoe prices across retailers",
  urls: ["https://nike.com/shoes", "https://zappos.com/nike"],
  schedule: "daily",
  status: "active"
}

Projects can be created via conversation or the API.

Schemas

A schema defines what data to extract from each page. Parser Run auto-detects schemas using LLM analysis.

Type	Description	Example
text	String values	Product name, brand
number	Decimal numbers	Price, rating
integer	Whole numbers	Review count, stock qty
boolean	True/false	In stock, on sale
date	ISO date string	Release date
array	List of values	Tags, sizes, colors

Jobs

A job is a single execution of a project's scraping pipeline. Jobs progress through stages:

1. discovery— Find all pages matching the target patterns
2. acquisition— Fetch HTML from each page
3. parsing— Extract data according to schema
4. complete— Data stored and ready to query

Records

Records are the extracted data rows stored in your project database.

Example: Product Record

{
  id: "rec_abc123",
  projectId: "nike-price-tracker",
  jobId: "job_xyz789",
  url: "https://nike.com/shoes/air-max-90",
  data: {
    product_name: "Nike Air Max 90",
    price: 129.99,
    brand: "Nike",
    rating: 4.7,
    review_count: 2341,
    in_stock: true
  },
  extractedAt: "2025-01-15T10:30:00Z"
}

HTML Archives

Every page fetch is archived for debugging and replay. This enables:

• Debugging — See exactly what the scraper saw
• Re-extraction — Update extraction logic without re-fetching
• Audit trail — Prove data provenance
• A/B testing — Compare extraction algorithms

Alerts & Webhooks

Configure notifications when your data changes:

Alert Configuration

{
  type: "price_drop",
  condition: {
    field: "price",
    operator: "decreased_by",
    value: 10,  // percent
  },
  notify: ["email", "webhook"],
  webhookUrl: "https://your-app.com/hooks/price-alert"
}

Next Steps

• Query data with the tRPC API
• Deploy your own instance with self-hosting