Introducing s-1: Sievly’s new document reading model
Workbench
Build document pipelines visually
Design, test and deploy document pipelines without code. Ship to production with a pipeline ID.
pipeline · invoices
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
Workbench
Design and test document pipelines
DEFINITION
Workbench is a visual space for designing document pipelines: drop in documents, tune options, watch the output update, then call the same pipeline by ID from your application.
WHO IT’S FOR
Engineers prototyping new flows, operations teams tuning schemas, and product managers checking quality before launch.
THE PROBLEM IT SOLVES
Pipelines often live as scripts only engineers can change. Workbench lets the people who know the documents adjust them safely.
Workbench in the platform
Using Workbench with the API
Configure in Workbench, deploy with a pipeline ID.
#1
When you are prototyping
Try a few documents and sketch a schema before writing integration code.
#2
When non-engineers tune a pipeline
Domain experts adjust chunking, OCR mode and fields in the interface.
#3
When you update production
Redeploy and your app picks up the new version by the same ID.
Where teams use Workbench
Pipelines that ship to production
Common jobs where Workbench is faster than writing code.
Prototype new workflows
Set up a pipeline against real files and hand the ID to engineering.
Let ops tune schemas
Domain experts update fields beside the source page.
QA pipeline output
Inspect citations, boxes and tables on the document.
Multi-step pipelines
Chain Parse, Split, Classify and Extract under one ID.
Versioned deploys
Every deploy is a version you can roll back.
Compare configurations
Run two configs side by side and keep the better one.
Build a pipeline on your own documents.
Why Workbench
Why teams use Workbench
01
Same engine as the API
Workbench and the API share one engine and one result.
02
Visual feedback
Citations and tables render on the source page.
03
Update without code changes
Redeploy and your app picks it up on the next call.
04
Usage-based
Included in every plan; pay only for pages processed.
How Workbench works
How Workbench works in four steps
■ STEP 01
Open Workbench
Sign in from your Sievly workspace.
workbench
■ STEP 02
Create a pipeline
Pick steps and name it. It gets a stable ID.
+ new pipeline
■ STEP 03
Upload and run
Drop in documents and tune options.
run · sample docs
■ STEP 04
Deploy
Publish a version and call it by ID.
pipeline_id
Built for production
Built on the same infrastructure as the API
Every deployed pipeline is a versioned, production-grade artifact that runs where the API runs.
Versioned pipelines and rollback
Each deploy creates a version you can inspect or roll back.
Role-based access and audit log
SSO, roles and audit logging on Enterprise.
Same SLAs as the API
Pipeline runs share the API’s infrastructure.
The rest of the platform
The endpoints around Workbench
/parse
Parse
Turn any document into structured, cited content.
Read more →
/extract
Extract
Pull the fields you define, with a citation on every value.
Read more →
/split
Split
Break long, multi-document files into logical parts.
Read more →
/classify
Classify
Label and route documents by type before they move on.
Read more →
/edit
Edit
Fill forms and update documents from plain language.
Read more →