Spiral Science and Technology Inc
Spiral provides a mobile-first automated visual inspection platform that runs on standard iPhones and iPads, eliminating the need for fixed camera infrastructure or extensive ML expertise. Operators capture training images on the shop floor, annotate components on a single frame, and the system automatically labels the rest of the dataset while synthesizing failure scenarios from one correctly assembled part. The platform returns real-time pass/fail verdicts with failure reasons tied to acceptance criteria, with first use cases live in as little as one week.
- Manufacturing / Industry 4.0
- Artificial Intelligence
- Industrial Automation
- Software Only
Funding
Funding not disclosed
Founders
Founder details are not available yet.
Product
Problem
Traditional industrial quality control relies on manual visual inspection, which is time-consuming, inconsistent, and requires trained operators at fixed stations. Automated inspection systems typically demand extensive labeled datasets, specialized hardware, and lengthy setup times, making them impractical for many manufacturing environments.
Solution
Spiral provides a mobile-first automated visual inspection platform that runs on standard iPhones and iPads. Operators capture training images on the shop floor, annotate components on a single frame, and the system automatically labels the rest of the dataset—no manual labeling or site visits required. The platform defines acceptance criteria by pointing the camera at a correctly assembled part, then synthesizes failure scenarios to learn exactly where differences occur. In production, the system returns real-time pass/fail verdicts with failure reasons tied to acceptance criteria, freeing operators from fixed stations to move with the work. The underlying ML pipeline can be configured end-to-end for specific assembly lines, with first use cases live in as little as one week.
Target Audience
Primary customers are manufacturers in aerospace and defense, wind energy, industrial piping, and marine equipment sectors that need flexible, on-site visual inspection without fixed camera infrastructure or extensive ML expertise.
Features
- No manual labeling: annotate components on a single frame and the system labels the rest of the dataset automatically
- Synthetic failure generation: defines acceptance from one correctly assembled part and synthesizes failure scenarios automatically
- On-device inference: models exported to CoreML for real-time pass/fail verdicts on iPhones and iPads
- Guided capture app: in-app prompts ensure consistent angle, lighting, and coverage per defect class
- Class balance checks: platform flags under-represented defect classes in real time during data collection
- Model types include defect classification, severity grading, anomaly detection, presence/placement verification, dimensional estimation, and text/marking recognition
- ERP/QMS integration: REST webhooks or direct connectors with configurable export formats (JSON, CSV, image package)