← All guides
Zero data egress
Upload a contract PDF and the tool extracts its commercial terms with a confidence score per field, then hands each one to you to accept, correct or reject.
Stage 1 of 2 · Per contract
Upload and AI extraction
Digital PDFs use pdf.js; scanned PDFs use Tesseract. Fields are found by a regex pre-pass, then NER and an on-device LLM for the rest.
What you see
- A PDF drop zone
- Extraction progress per field
- Populating field cards
What you do
- Drop a contract PDF
- Wait for parse and extraction
- Watch the field cards populate
What you get
- The key commercial terms — payment terms, discount schedules, notice periods, expiry and auto-renewal dates, volume thresholds, contract value, supplier name, liability cap and governing law
- A confidence score per field
- Ready for your review
Stage 2 of 2 · Ongoing
Review and summary
Every extraction is yours to accept, correct or reject. Only your accepted values ship downstream.
What you see
- A contract summary panel
- Extracted-field cards with confidence
- A source-passage jump for each field
What you do
- Review each field one at a time
- Correct any low-confidence extractions
- Click Accept per field
What you get
- Structured contract terms stored locally
- Fed to LiquidityAssist for early-payment ROI
- Fed to the cross-tool alerts engine
Heads up. Scope is deliberately narrow: this extracts commercial terms only. It is not a contract lifecycle management tool — no authoring, obligation tracking or approvals.