Takeoff & Estimating
Measure once.
Bid with confidence.
AI-powered takeoff for construction. Upload plans, detect rooms, build estimates — cloud-ready out of the box, with optional self-hosted AI for teams that want full control.
What's in the box.
Purpose-built tools for construction estimators. Not a generic spreadsheet with a plan viewer bolted on.
Digital Plan Takeoff
Upload plan sets. Navigate multi-page drawings, calibrate scale, measure directly on plans.
Precision Measurement
Area polygons, linear measurements, vertex snapping. Automatic SF, LF, and CY calculations.
AI-Powered Detection
One click identifies rooms and boundaries. Accept, adjust, or refine — every correction trains the model.
Assembly-Based Costing
Build assemblies layer by layer. Attach to conditions for instant quantities and cost rollups.
Multi-Tenant Platform
Teams, projects, org-specific pricing. Role-based access and audit logging from day one.
Estimate to Bid
Measured conditions to priced line items to total bid. Markup, OH&P, and export.
From Takeoff to Bid
Every square foot has a price.
From measured condition to priced line item. Assemblies, waste factors, and markup — the complete estimating workflow in one place.
- Assembly-based costing with layer-by-layer material rollups
- Automatic SF, LF, and cubic yard calculations at architectural scale
- Markup, OH&P, and bid export — ready for the GC
AI That Learns From You
Correct it once. It never forgets.
Built on SAM 3, Meta's open-source segmentation model. Every polygon you adjust becomes training data for a model fine-tuned on your plan types.
- Computer vision detects rooms and regions from architectural drawings
- Every estimator correction becomes training data — the model learns your plan types
- Continuous improvement: detect, correct, retrain, detect better
Your Data, Your Rules
Organizations control their training data. Keep corrections private to your org, or opt in to share with the platform — improving detection for everyone while maintaining full ownership of your data.
Curate AI predictions. Accept, flag, or reject — every decision improves the model.
How It Works
Orchestrated, not monolithic.
Barker coordinates the workflow. AI agents — in the cloud or on your hardware — process plans by reading directly from storage. No bottleneck, no round-trips through the app.
Agents read plans and write results directly to storage — no data passes through the app. Faster processing, lower latency, and your plans never leave your control.
Optional Self-Hosted AI
Go further with your own GPU.
AI detection works without any special hardware. But if you want faster inference, zero per-query costs, and full control over your models — plug in a GPU you already own.
Optional: RTX 5070 connected. Inference in ~300ms.
Bring Your Own GPU (Optional)
AI detection works out of the box. Want faster inference and zero per-query costs? Connect your own hardware — workstations, servers, gaming rigs.
Iterative Learning Loop
Every polygon correction feeds back into model training. Weekly updates reflect real estimator expertise.
Open-Source Foundation
Built on SAM 3, Meta's open-source segmentation model. You own your training data, your fine-tuned models, and your edge.
Faster Takeoffs
AI detection eliminates hours of manual tracing.
Team Ready
Multi-tenant from day one. Org pricing, roles, audit trails.
Gets Smarter Over Time
Your corrections compound into a durable advantage.