Introducing s-1: Sievly’s new document reading model
All posts
Announcements
Introducing s-1, our new document reading model
s-1 reads PDFs, scans and spreadsheets the way a careful analyst does, and returns structured data with a citation for every value. It is available in preview today.
Mara Lindqvist
·
Co-founder
sievly
Introducing s-1
Today we are opening the preview of s-1, the reading model that now sits underneath Parse, Extract, Split and Edit. It replaces a stack of stitched-together tools with a single model that understands layout, tables, figures and handwriting in one pass.
Why we built a new model
Most document pipelines break in the same places: merged table cells, rotated scans, footnotes that change the meaning of a number, and charts that hold the only copy of a value. Teams were spending more time patching those edge cases than building their product. s-1 was trained to read those pages the way people do, starting from the layout and working down to every field.
What changes for you
One flat per-page price for every document type
Citations and bounding boxes on every extracted value
Better handling of nested tables, handwriting and mixed-language files
The same API calls you already use, with a new model option
How to try it
Open Workbench, drop in a few of your hardest documents and switch the model to s-1. Your existing pipelines keep running on the current model until you choose to move them, so you can compare results side by side before you switch.
More from the blog
sievly
Reading order
Research
Research note: reading order in complex layouts
Multi-column pages, sidebars and callouts confuse simple reading order. We share what we learned teaching s-1 to follow the page the way a person would.
sievly
Quarryline × Sievly
Case Studies
How Quarryline turns research notes into a searchable knowledge base
Quarryline’s data team uses Parse and Extract to turn thousands of analyst notes, filings and spreadsheets into clean, cited records their AI assistant can trust.
sievly
Layout-first chunking
Engineering
Layout-first chunking: why we split by structure, not tokens
Fixed-size chunks cut tables in half and separate headers from their rows. Here is how Sievly decides where one idea ends and the next begins.