One person, one project.
- 1 seat
- 1 project · 1 repository
- 2 GB storage
- Limited daily usage
Vero is the vision AI agent inside VisionRepo, a computer vision platform for your images and video. Ask it to find, label, train, test and report. You approve every step.
Wafers, boards, panels, fabric. Any visual data, labeled the same way.
scratch
missing part
hotspotbypass diode
thread pull
crack
particleedge chip
hotspot
missing part
hole
particleWhat changes when your visual data, your models and Vero share one AI vision platform.


Catch defects with models trained and tested on your own parts, boards and panels.

Every AI label is a draft. Every review is recorded. Every version is kept.

Accuracy for every model, and tests on new data before anything goes live.

Ask Vero for labels, counts, charts and scheduled reports.
Five steps in the real app. Click a step, or let it play.
Speak to your data in plain words. Vero finds images, labels folders, trains and compares models, and builds dashboards. Every change waits for your OK.
Built on the same platform. No model to train, no code to write.
Drone thermal and RGB photos in. Critical, Mild and Low PDF reports out, with a panel code for every anomaly.
Explore SOLAR-1Inspect boards against up to three golden templates. Pass or Defective per board, with first-pass yield.
Explore PCBA-1Folders, metadata and tags, filters you can combine, bulk actions across the whole repository, and versions every model points to.
One person, one project.
A small team labeling and training.
Teams labeling together.
Inspection on the line, on your own servers if you need it.
The vision AI agent built into VisionRepo. Speak to your data, and let Vero label it and train your model. Ask it to label, train, chart or explain; it shows a plan and waits for your approval. Every action is also a button.
A computer vision platform, and a vision AI platform, by Averroes.ai, with a vision AI agent, Vero, built in. Store images and video, label them, and train models you can test, compare and call over an API, by hand or by asking Vero. SOLAR-1 and PCBA-1 turn the same engine into solar reports and board results.
You land in an empty workspace with Create Your First Repo. Upload images and video, or Try a Sample, then label, train and test. The upload dialog shows those four steps.
Training starts at 10 labeled images; aim for 20 to 50 per class. Anomaly detection (Ultra) needs only images of good parts.
Yes: $0, no card, 1 seat, 1 project, 2 GB. Pro is $20/mo, Ultra $200/mo.
VisionRepo runs in the cloud and serves models over an API. Inline, on-prem and air-gapped inspection is Averroes production inspection.
Free plan, no credit card. Bring your own images or start from a sample.