From plan PDF to a sendable bid in 30 minutes.
Stop manually counting symbols on plan sheets. Stop losing bids to contractors who can turn around estimates in hours, not days.
Accuracy that's been pressure-tested on real plans
Accuracy versus manually verified takeoffs on real grading plans
From plan PDF upload to a sendable bid, end to end
Lines of extraction logic, calibrated against real grading and site bids
Grademark runs a frontier multi-model ensemble with a custom keynote legend extractor that handles the OCR artifacts most plans throw. keynotes drawn inside circles, hexagons, oddly placed callouts. Where competitors give up and force a human into the loop, Grademark uses positional inference and falls back gracefully.
AI takeoff, estimator judgment
Grademark doesn't replace your estimator. It removes the drudgery so they can focus on judgment. The flow:
1. Auto-extract
Plan in, scope out
The AI reads the keynote legend, measures the linework, and produces a confidence-scored takeoff per scope item.
2. Review & override
Estimator stays in the loop
Every line item is editable. Override quantities, swap unit costs, add scope the AI missed. Confidence below 80% is flagged for review.
3. Export
PDF, Excel, ERP
Branded bid PDF, Excel template export, or push directly into the Grademark ERP project pipeline.
What the takeoff actually looks like
Real Grademark output. Line-by-line scope items extracted from the plan, with quantities, unit costs, line totals, and a confidence score for each. Below 80% gets flagged for estimator review.
Quick Bid · grading_plan.pdf
Status: ready
| Scope item | Qty | Unit | Unit cost | Total | Conf. |
|---|---|---|---|---|---|
| Mass excavation, cut | 8,420 | CY | $4.20 | $35,364 | 92% |
| Structural fill | 3,210 | CY | $6.50 | $20,865 | 88% |
| Storm drain, 18" RCP | 1,212 | LF | $48.00 | $58,176 | 95% |
| Catch basin | 12 | EA | $1,800 | $21,600 | 100% |
| SWPPP, silt fence | 2,140 | LF | $4.20 | $8,988 | 85% |
| Retaining wall, 4' CMU | 420 | LF | $185 | $77,700 | 78% |
| Subtotal | $222,693 | ||||
Illustrative output. Your bids use your unit costs and your scope.
Manual takeoff. Estimator overrides.
The AI takeoff is a draft. Estimators get the final word. From the worksheet you can:
- Override quantities and unit costsClick any cell. Type the corrected value. The line total and bid total update live.
- Add scope the AI missedPick the page, draw the measurement on the plan, set quantity and unit. Pushed straight into the worksheet.
- Toggle scope in or outItem is excluded by the GC? Uncheck it. The total adjusts and the bid PDF prints clean.
- Confidence flagsAnything below 80% is highlighted. You see exactly where the AI was unsure and what to verify.
Manual takeoff · estimator review
How Grademark compares
| Feature | Grademark | Manual Excel | Bluebeam | HCSS HeavyBid |
|---|---|---|---|---|
| Plan PDF auto-extraction | ✓ | ✗ | Partial | ✗ |
| Calibrated to your prices | Via Grademark ERP | Manual | ✗ | ✓ (manual setup) |
| Time per bid | 30 min | 4–10 hrs | 2–4 hrs | 1–3 hrs |
| QuickBooks Desktop sync | Via ERP (shipping shortly) | ✗ | ✗ | Add-on |
| Starting price | $99/mo or $5K once | N/A | $260 once | $5,000+/yr |
| Site work focus | ✓ | N/A | General | Heavy civil |
Bid tool FAQ
- What plan formats do you support?
- Any PDF. Vector or raster. We OCR scanned drawings.
- What if the AI extraction is wrong?
- Every line item is editable. Confidence scores flag anything under 80% for review.
- Can I keep using my current bid format?
- Yes. We export to your branded PDF or Excel template.
- Is my plan data private?
- Yes. We don't train on your plans. Single-tenant data isolation.