Introducing s-1: Sievlyโs new document reading model
Turn every page
into data.
The document platform for AI teams that need dependable accuracy on real-world files. One API for every document task.
Helping everyone from early-stage startups to global enterprises unlock their data.
Quarryline
Brindlewood
nimbleton
brightwick
vantrel
PLINTH & CO
OAKMERE
ARROWFEN
Halvory
pebblegate
Mirrowind
quillstack
Whatโs new
Introducing:
s-1 Model
s-1 is our most capable reading model yet. It replaces stitched-together, multi-tool pipelines with a single model and one flat per-page price. Available in preview today, so you can try it on your own documents in Workbench.
0:00 / 1:24
Product
Free your data from documents.
A set of flexible APIs for every document job.
Python
Node
$ pip install sievly
Use cases
Trusted where every field counts.
Chosen by finance, healthcare and legal teams, where one wrong value can undo a whole workflow.
Use cases
Finance
Pull key figures from investor decks, oversized spreadsheets, dense pitch materials and regulatory filings. Nested tables, charts and statements come back intact, down to the last decimal.
Sievly read files we had given up on because the tables were too complex. It is the first document tool that has actually held up for us.
Head of Data Engineering, Global Asset Manager
How it works
Built to read the way people do.
Sievlyโs multi-pass system pairs OCR with vision language models for accuracy you can rely on.
See the whole document
s-1 reads every element of a page at once, so text, tables, figures and styling keep their meaning.
Make sense of complex structure
One unified reading model handles dense tables, unusual layouts, annotations and information that depends on formatting.
Return structured, cited output
Each response maps the documentโs structure and ties every value to its source, so the next step can use it right away.
From
PDF to JSON
in a single API call
Features
One platform for every document workflow.
A wide feature set that covers the long tail of your processing needs, ready out of the box.
Layout extraction
Read sprawling tables and handwritten forms alike, with page positions and structure kept intact.
File type support
One endpoint accepts PDFs, scans, spreadsheets, decks and many other file types.
Language support
Multilingual parsing across many scripts, including mixed-language documents.
LLM ready
Smart chunking, embedding prep and everything else you need to make your data ready for models.
Image support
OCR for scanned pages, faxes and handwriting, accurate even on low-quality inputs.
Smart chunking
Figure summaries
Chart extraction
Page rotation
Embedding prep
Smart chunking
Figure summaries
Chart extraction
Page rotation
Embedding prep
+ many more features
Everything else you need to make your data ready for models.
Enterprise-ready
Security, reliability and support for teams running document AI in production.
High-availability uptime
Proven systems that stay steady as your page volume climbs.
Enterprise support and SLAs
Engineers who work alongside your team, plus service terms shaped around how you operate.
Built for regulated teams
Safeguards designed for confidential files and regulated workloads. Read our security policies.
Deploy in your environment
Keep every document on your own servers when security reviews, regulators or data location rules call for it.
northpine
Calverd
Ostrava Bay
โ Meridane
โข larkfold
Questions, answered.
Do you offer trials? How do I get an API key?
You can get API access right away on our Standard plan, which is pay-as-you-go. If you would like to trial the Growth or Enterprise plans, fill out the contact form and we will get you set up.