Sign in Sign up
Loading viewer…
Loading run data…

Generating schema…

Analyzing your example documents to scaffold fields.

Schema Builder

Property key Type Optional

Schema exceeds the 32 KB limit — shorten descriptions or remove fields.

Read-only preview of the schema sent to the API.

Extraction rules

Optional natural-language guidance for the extraction model, applied on top of the schema.

Advanced options

stable is the production default; nightly is the latest preview build.

Parser configuration

structure for OCR-backed extraction, understand for deeper document analysis, or agentic for VLM-augmented extraction.

Auto-detect runs offline language detection; or pick a specific language. For 130+ languages, use the API.

The most distinct writing systems (scripts — e.g. Latin, Cyrillic, Han) the detector may find on a page.

The most distinct languages the detector may find. A single script often covers many languages (Latin spans English, German, French…).

Running extraction…

Extracting fields from

Larger documents can take a little longer.

Processor info

ID
Type
Created

Published versions

Published versions are immutable. Select a published version to view its configuration and run it.

Filters

Status

Updating runs…

New processor

Publish ?

Publishing is permanent — a published version is immutable and can be run via the API.

/500

Edit name

has been deleted and can no longer be opened. The run keeps its executed configuration as a receipt.

Import JSON Schema