shenshen Start free

SYS-001 · Recognition × Parsing

Your team has better things to do than retype images

shenshen reads the images your business runs on — crooked photos, glossy scans, shots of screens — and returns the exact fields your systems need. Your people stop proofreading and get back to the work only they can do.

Start free — no card required Talk to us about your workflow

Free to start · No card required · Early pricing · locked in for life

Photo in → spreadsheet row out

Reads the documents your team already has Invoices Business cards Screenshots Contracts
FIG·02 — Pipeline

Two engines. One spreadsheet row.

Recognition reads the photo. Parsing labels what it found. Everything downstream gets clean fields instead of text to re-format.

Engine A · Vision
Recognize
  • A1Deskew — rotation & perspective corrected
  • A2De-glare — glossy & screen reflections removed
  • A3Sharpen — low-light fine print recovered
Engine B · Parsing
Label
  • B1Type — invoice, card, table, contract
  • B2Fields — vendor, number, total, dates…
  • B3Validate — totals & dates checked
Two engines, merged into one output
VendorInvoice #Line itemsTaxTotalDue date
FIG·03 — Document types

Fields, not a wall of text

Text you still have to re-format. Fields you can sort, sum, validate, and send. That's the difference.

Invoices

Fields returned

vendor, invoice number, line items, tax, total, due date

Business cards

Fields returned

name, title, company, email, phone

Screenshots

Fields returned

table → clean spreadsheet rows

Contracts

Fields returned

parties, effective date, payment terms, key clauses

FIG·04 — Acquisition conditions

Built for the photos people actually take

We test against the documents your team really photographs — not clean scans.

Crooked, not broken

A receipt shot at an angle on a desk gets deskewed before reading. Rotation and perspective are corrected first, so lines stay straight.

Glare and glossy screens

A laminated card in bright light, or a phone screen with a reflection — the glare pass removes it before the recognition pass starts.

Low light, fine print

A dim photo of a contract is upscaled and sharpened before parsing, so fine print isn't the reason a clause goes missing.

FIG·05 — Integration

Drops into the workflow you already have

Output goes where your work already happens.

CSV & Google Sheets

Rows land in your spreadsheet — not a PDF you can't edit. Paste or import and the numbers are already separated.

API access

Send shenshen a photo, receive a structured object back. Build the exact workflow your team needs.

No card to start

Try it on your own documents free. No card, no sales call, no setup meeting.

FIG·06 — Pricing

Free to start. Plans that scale with your team.

Start free on your real images. When your team is ready, we'll build the plan around your volume.

Free to start
Freeno card required

Test shenshen on your real images. Enough volume to see how it handles your messiest photos.

  • Recognition + parsing for all four document types
  • Manual field review and edit before export
  • Export to CSV & Google Sheets
  • Community support
Start free
Limited seats Founding member
Early pricinglocked in for life

Early adopter pricing locked in for life. Seats are limited — contact us for current pricing and availability.

  • Everything in Free, unlimited
  • Unlimited documents
  • API access for automation
  • Founding member badge
  • Priority support
Ask about founding seats

Early pricing · locked in for life

Early access

Get your team's hours back

Leave your email and we'll let you know the moment founding slots open. Free to start — no card required.

We'll only email you about shenshen. Unsubscribe anytime.

FIG·07 — Reference

Questions from teams evaluating a tool

What exactly does shenshen extract?

It depends on the document type. Invoices yield vendor, invoice number, line items, tax, total, and due date. Business cards yield name, title, company, email, and phone. Screenshots of tables come out as spreadsheet rows, and contracts yield parties, effective date, payment terms, and key clauses.

How accurate is the recognition?

On a clear, well-lit invoice photo, totals and dates come out as numbers you can sum and sort. On a receipt shot at an angle or a card with glare on a glossy surface, shenshen corrects the photo first — deskew and de-glare happen before reading — which is where a lot of older tools drop out.

Where does our data go?

Your documents are processed by our service under our privacy policy. We don't use your documents to train models for other customers, and you can delete a document at any time from your workspace.

How do we get the data out?

CSV and Google Sheets export for humans, and an API that returns a structured object for automation. The fields are labeled, so a total is a number and a due date is a date.

Is there really a free plan?

Yes. Free to start, no card required. You can test shenshen on your own documents before paying anything.

What does pricing look like?

Free to start, no card required. Paid plans scale with your team's volume. Founding member seats come with early adopter pricing locked in for life. Contact us for current pricing.