
1. Technology overview
GAZIRU Inc. was founded in April 2020 and is based in Setagaya-ku, Tokyo. It develops image-based AI for visual inspection, non-destructive testing and individual item identification in manufacturing.
GAZIRU.eye is anomaly detection software that trains a model from a few dozen normal images, with no defect samples needed. Industrial cameras capture images and inference runs on an inspection server in under one second. Users can swap AI engines or run several in parallel, view defect heatmaps, and connect through a REST API to production systems. An optional link to GAZIRU.z ties results to individual items for traceability.
2. Highlighted features & specifications
- Normal-image training: builds a model from a few dozen normal images.
- Swappable engines: choose, change or run several AI engines in parallel per product.
- Defect heatmaps: visualize anomaly locations for audits.
- Fast inference: under 1 second per image on an inspection server.
- REST API: connects to production management systems and inspection equipment.
- Traceability link: optional GAZIRU.z integration ties results to individual items.
3. Potential application
Detecting scratches, dents and paint unevenness on metal parts at a Vietnamese auto-parts supplier. Flagging foreign matter on food production lines. Inspecting weld or casting internal defects with ultrasonic images.
4. Benefits
No defect samples are needed, which suits high-mix, low-volume production. Heatmaps explain where anomalies are for quality audits and customer discussions. Inference completes in under one second. Traceability is possible through the optional individual identification link.
5. Market trends & opportunities
Factories adopting AI visual inspection often lack enough defect images, so normal-image learning lowers the barrier. Vietnamese electronics, metalworking and automotive suppliers under customer quality audits need explainable inspection results and traceability.
6. Other information
Founded April 2020, Tokyo. Official launch January 2026. Applications include scratches, dents, dirt, foreign matter, paint unevenness and non-destructive testing in automotive parts, electronics, metalworking and food. Line integration is typically custom-built per customer.
One platform connecting startups, corporations, research institutions, regions and investors











Playbook & Handbook · February 6, 2026
Winning Accelerator Playbook 2026 by BambuUP
RETAIL E-COMMERCE: AI AND REAL-TIME DATA ARE TRANSFORMING HOW RETAIL BUSINESSES OPERATE








































