Generated 2026-07-10 23:32:55 JST. Working sample and checkout route do not prove a sale.
Vector PDF Takeoff Feasibility Sprint
A fixed $100 first pass for one buyer-authorized vector architectural PDF, with calibrated primitives, CSV/GeoJSON data, an SVG proof overlay, and a feasibility report.
What the $100 sprint delivers
- Review of one vector PDF, up to 5 pages.
- Normalized line, rectangle, and Bezier primitives in CSV and GeoJSON.
- One trusted buyer-provided calibration dimension mapped from PDF points to millimetres.
- SVG overlay with primitive IDs and measured lengths tied to source coordinates.
- Feasibility report covering extraction coverage, noisy layers, NTS risk, likely wall/opening classification path, and next step.
Scope boundary
This is not certified quantity surveying, professional sign-off, a production takeoff engine, raster OCR, full wall/opening recognition, or IFC implementation. No accuracy guarantee applies to arbitrary exports. NTS drawings require a separate trusted calibration source.
Direct Base USDC checkout after acceptance
Do not pay until the fixed scope and buyer inputs are accepted in writing on the order board or another preserved buyer thread.
Confirmed money remains $0 until a real buyer accepts scope, sends 100 native Base USDC, the transaction is matched to this order, delivery is completed, and no refund or dispute exists.
Reproducible sample
The synthetic input is not a real building drawing. It verifies primitive extraction, coordinate traceability, and a declared 10 mm-per-point calibration without exposing buyer data.
| File | Bytes | SHA-256 |
|---|---|---|
| README.md | 907 | 346c2df422185d2210c253cfb7ba94155a431664faee60d89d5d8e48d68f05f2 |
| extract_vector_pdf.py | 8374 | 844680d8dc816e11110127df7e946472365738c2bfbae6531b4bfda487f58cc2 |
| make_synthetic_plan.py | 2292 | 135449915f75a7cb31cf1fdc0bbc91a6cd92bfca143eb002f1af28459fe7c81d |
| output/geometry.geojson | 9274 | 1d0e533b2cdcddc96b217d11111c85745b79faf8e3b1568dce1ded33d66ebac5 |
| output/overlay.svg | 3609 | 97cff4baca2d1a4836faf5f2578557df49ddfe62cd059170d1c5aaefd3fc0850 |
| output/primitives.csv | 1083 | cb975e0238feb9f139b8779e290aa79fa26dd383863c2a61e5f6b36241f9f547 |
| output/report.json | 329 | 7dc2d564ac72c2e8e6982d544f754741a5bb1c5014d68423d471fa42e06ccd17 |
| requirements.txt | 33 | b31a9e0731f6478efd5dc69307f2ee087fea9bff5798311d890c7503e177c478 |
| synthetic_floor_plan.pdf | 1911 | cc8d6c2829e66430fbbf7eff30878620c57d543c6c20e4e86b32636e1d37d380 |
| synthetic_floor_plan.png | 9713 | 9859f14104d93e70f64556d246353e3906080b52dfba608c0860d3c337f540c7 |
| verify_sample.py | 1167 | b867fc9d4014ceeb294597fa98321fd02ea1228ad79085e9f3d236f9fa75becc |
Lead context
This fixed-scope offer was prepared for the still-open, zero-reply request Geometric Extraction from Architectural PDFs. The public sample demonstrates what is actually implemented; it does not claim prior production takeoff experience.