Planwise: AI electrical takeoff in minutes from residential plans
ScalaCode built Planwise, a web SaaS that uses OCR and computer vision to auto-count symbols on residential electrical plans and produce accurate material takeoffs in minutes, so electricians can quote faster and spend far less time counting symbols by hand.
An emerging construction technology company
The client wanted to automate electrical takeoff and material estimation from residential drawings, eliminating manual symbol counting so electricians could generate accurate material lists and quoting inputs within minutes.
Why automated takeoff is hard
Inconsistent electrical legends
Every builder and architect draws legends differently and there's no standard format, so rule-based detection couldn't be relied on.
Complex, noisy plan layouts
Symbols can be rotated or mirrored, and annotations, stamps and dense notes overlap them. Each of those makes a miscount more likely.
Accuracy with manual control
Electricians still had to review and correct the results, so the automation needed to leave room for a person to check it.
AI processing costs at scale
Running OCR and computer vision over large drawings costs money, and that cost had to stay under control as usage grew.
An AI pipeline that reads each plan's own legend
The pipeline turns a complicated drawing into takeoff data that's accurate and can be edited, with very little manual work.
- 01
OCR and legend extraction
OCR reads the legend definitions and annotations off each drawing automatically.
- 02
Context-aware symbol detection
The vision models work from the legend pulled out of that drawing, so each symbol is classified the way that drawing defines it.
- 03
AI fallback for noisy plans
When a plan is noisy, a second reasoning layer deals with overlapping annotations and stamps, so nobody has to clean the plan up by hand.
- 04
Visual validation
Detections are drawn over the plan, so an electrician can check edge cases and fix them right there.
- 05
Configurable subscriptions
Plan limits, features and usage thresholds sit in configuration, and changing the pricing doesn't need a code release.
Multi-hour takeoffs now take minutes
Electricians upload multi-page plans, the pipeline reads the legend, counts lighting, power, fan and data points, and produces structured reports they can check and edit.
takeoff cycle time
manual edit control
Inside Planwise
- Analysis results with symbol counts
- Processing summary, ready to download
What Planwise runs on
React.js, Tailwind CSS
Node.js, PostgreSQL
Google Vision API
AWS EC2, S3, CloudWatch, Stripe
Plan a build like Planwise for your business
- 30 minutes with the ScalaCode team
- What is realistic, what it costs and how long it takes
- You decide afterwards whether to build with us
