Schema Source by Tabstack - Turn any URL into a JSON Schema, Zod, or Pydantic model

by
Get the JSON Schema for any URL. Paste a link and get a ready-to-use schema as JSON Schema, a Zod object, or a Pydantic model. Copy it straight into your extraction or validation code. Schema Source learns each site's URL patterns, so results come back fast and stay consistent across pages of the same shape. Every result is also an API: request the same URL with JSON content negotiation and get the cached schema back, no key required. Free, by the team behind Tabstack—backed by Mozilla.

Add a comment

Replies

Best

This team delivers.

is a powerful web content extraction and transformation toolkit designed specifically for AI agent builders. It's schema-based. You define a JSON schema for the fields you want, and every call reads the page and maps it to that schema. So you maintain the schema, not the scraping logic.

When they , the community had many questions on this schema-first approach.

Today, the team introduces - a free utility to help you quickly build your first schema and get started. Enter any URL, get a structured JSON schema. That's it!

You have the API. You have the tools. You now have the schema. What will you scrape today?

 Nice! This makes sense now. Stumbled upon Schema Source yesterday. So, Schema Source + Tabstack is the combo. Congrats on launch!

 exactly - thanks for the support!

👋 Product Hunt!

We're launching our free tool, Schema Source. Paste any URL and it generates a ready-to-use JSON Schema for that page's data, as JSON Schema, Zod, or Pydantic.

Generating your own is great when the page is unusual or specific. But a lot of extraction work is the same handful of shapes over and over: a job posting, a product listing, a company profile, a real estate listing. So here's another way to do it.

We also open-sourced 46 pre-defined schemas on GitHub, across 10 industry categories: real estate, jobs, e-commerce, finance, healthcare, dev tooling, gov records, travel, b2b intel, and social. Grab the file, point Tabstack at a URL, get data back.

Grab one or generate your own, whichever fits.

 OSS ftw!!