# OCR-API.xyz

> Document parsing API. POST a PDF, get back clean Markdown, an ordered layout JSON with bounding boxes, token-level OCR JSON, and every extracted figure as a signed image URL. One endpoint, signed webhooks, hash-based caching. $0.01 per page. 100 free credits on signup — no credit card required.

OCR-API.xyz parses PDFs into structured outputs:

- **Markdown** — full document text, good for chunking or search.
- **Layout JSON (`content_list`)** — ordered blocks `{ type, text, text_level, bbox, page_idx }`.
- **OCR token JSON (`middle`)** — every word positioned with a bbox.
- **Images** — signed URLs to every extracted figure / diagram.

Capabilities: Tables, Forms, Images, Graphs, Equations · Layout, Bounding boxes, Structured schema, Extraction citations · Multilingual OCR, Automatic page rotation, Table summarization, Enrichment VLM, Spreadsheet table clustering, Document splitting, Custom webhooks.

Currently supported file format: **PDF**.

## Pages

- [Landing page](https://ocr-api.xyz/index.md): product overview, capabilities, pricing.
- [API documentation](https://ocr-api.xyz/docs.md): full reference for `POST /v1/extract`, `GET /v1/jobs/{id}`, webhooks, errors.

## Quick start

- Base URL: `https://api.ocr-api.xyz/v1`
- Auth: `Authorization: Bearer pk_live_…`
- Submit: `POST /v1/extract` with multipart `file` (PDF) and optional `webhook_url` + `webhook_secret`.
- Poll: `GET /v1/jobs/{id}` until `status === "done"`.
- Billing: 1 credit = 1 page = $0.01. 100 free credits on signup. Auto-recharge $50 when a card is attached.