← All portfolio work

04 · Consumer Legal AI · Since 2024

Should I Sign AI

Understand any legal document in seconds, before you sign it.

Built for: Individuals and small businesses signing contracts without a lawyer on call

Product preview

shouldisign.ai
Drop a lease, NDA, or offer letter
Overall risk (as tenant)Medium
Early-termination penalty
Security deposit terms
Standard renewal clause

Illustrative interface, rebuilt for this page - not a live screenshot.

What this is

Most people who need to understand a contract can't justify calling a lawyer for a five-minute question. Should I Sign AI takes any uploaded document - a lease, an NDA, an employment offer, a court order - and returns a plain-English risk read in seconds, from the perspective of whichever party you actually are.

The challenge

Legal risk analysis had to be fast enough to use in the moment someone is deciding whether to sign, cheap enough to use on a low-stakes document, and accurate enough to trust on a high-stakes one - three requirements that usually trade off against each other.

How we built it

01

Perspective before analysis

The first question isn't 'what does this contract say', it's 'which party are you' - a lease reads differently for a tenant than a landlord, and the risk report reflects that from the first sentence.

02

Two speeds, one interface

A fast mode gives an immediate first read; a deep-analysis mode runs a more thorough pass for documents where the extra minute is worth it - the same interface, a mode toggle, not two products.

03

OCR first, so no document is excluded

Scanned PDFs and photographed documents run through OCR before analysis, so a phone photo of a lease is treated the same as a clean digital copy.

04

Reports that can be shared, not just read once

Every analysis can generate a shareable public report link, so a tenant can send the risk read to a partner or a small-business owner can loop in a co-founder without re-running the analysis.

What it does

Document upload and OCR

Accepts PDFs, scans, and photos; Tesseract-based OCR normalises anything image-based before analysis begins.

Perspective-aware risk report

Clause-by-clause risk flags generated from the chosen party's point of view, not a generic neutral summary.

My Documents library and reminders

A personal library of past analyses with reminders for key dates found inside the documents.

Compare

A side-by-side view for comparing two versions of a document or two competing offers.

Teams and API access

Multi-seat accounts for small businesses and API keys for programmatic access, alongside the individual consumer product.

Content hub

A blog explaining common clause types and risk patterns, so the product also teaches, not just flags.

Tech stack

React 18 + Vite
The core application shell, chosen for fast iteration on a high-traffic consumer product.
Tesseract.js
In-browser and server-side OCR so scanned and photographed documents get the same analysis quality as digital ones.
pdf.js, mammoth, docx
Parses PDF, Word, and plain-text documents into structured text before the risk model ever sees them.
Supabase
Auth, document storage, and the shareable public-report layer.

Outcome

Document types supported
Contracts, NDAs, leases, offers, orders, POAs
Analysis modes
Fast and deep
Access surfaces
Web, teams, and API