No file selected
Drop your documents here
PDF, PNG, JPEG, TIFF or WebP · multiple files supported
Browse files
text
table
recovered image text
uncovered (fallback)
header/footer
Parse
›
Extract
›
Classify
›
Split
Health
Config
Results
Chat
Run parser
Run all files
Run new files
Layout detection
Detection model
How regions (text, tables, figures) are found on each page.
Standard + forms (recommended)
Standard
High accuracy (slower)
Fast (less accurate)
Split form grids into separate fields
Helps on boxed forms (e.g. ACORD) where sections share one grid.
Text extraction
Capture mode
Full text also recovers words that layout detection missed.
Full text (recommended)
Layout regions only
Simple reading order
Rendering
Render quality (DPI)
200 is a good default · 300 for small print (slower).
100
150
200
300
Run the parser to see results.
Formatted
Blocks
Raw
JSON
Download .md
Download .json
Download plot (PNG)
Download all
Batch summary
Schema builder
Suggest schema
Fields to pull from each document: key, type, and a description that guides the model.
Add property
Delete all
Extraction model
Runs on OpenRouter after the document is parsed by the OCR service.
Loading models…
Run extraction to see structured fields.
Fields
JSON
Download extraction .json
Classification types
Document types the classifier can choose from; the description guides the model.
+ Add classification
Classification model
Runs on OpenRouter after the document is parsed by the OCR service.
Loading models…
Run classification to see the document type.
Split types
Sub-document types for splitting multi-page files; the description guides the model.
+ Add split type
Split model
Runs on OpenRouter after the document is parsed by the OCR service.
Loading models…
Run the splitter to see sub-documents.
Loading models…
Clear
Ask anything about the parsed documents —
their extracted content is sent as context.
Send