VisionQ

Software that looks at things and refuses to guess.

A verdict arrives with its measurement. A count arrives with the range it can be trusted in. A model arrives with the hash that identifies it. Five products, each shown with the status its own repository states, on the date it was read.

See the products Read the status ledger

FindCircle · r within 20–120 px
VisionAI seed program

Five products, one discipline

Each one is presented as a datasheet: what it runs on, what it governs, and what it hands off.

Vision suite

In development means working software being built on. Design means schemas and scaffolding with nothing qualified.

VisionAI

In development

Inspection programs that return a verdict with its measurement.

Runs on An edge device or site node on the plant network
Governs Programs, jobs, vision tasks and a single-active-operator lock
Hands off to Verdicts and measurement overlays, with run history
Checked 2026-09-01 Read the sheet

VisionML

In development

Model training that ships a manifest saying exactly what it made.

Runs on A training environment, separate from anything that runs inference
Governs Datasets, training runs, evaluation, export and the manifest contract
Hands off to An ONNX artifact and a manifest.json, pinned by sha256
Checked 2026-09-01 Read the sheet

VisionEdgeOS

Design

A signed product image for inspection on industrial edge devices.

Runs on Industrial edge devices, on a supported upstream Linux
Governs Hardware profiles, release manifests, SBOM and provenance, revocation and commissioning records
Hands off to A verified image and a commissioning record, and version observations to VisionFleet
Checked 2026-09-01 Read the sheet

VisionFleet

Design

Where every device is, what it runs, and how far a rollout has got.

Runs on A central service, with the console in a browser
Governs Device assignment, exact component inventory, governed groups and target snapshots
Hands off to Desired-versus-observed drift, rollout progress and incidents
Checked 2026-09-01 Read the sheet

Laboratory

Digital Labs

In development

Session, capture and review in one thread, on iPhone, iPad and Mac.

Runs on iPhone, iPad and Mac, local-first
Governs Bench sessions, version-pinned protocols, captures, deviations and review state
Hands off to A reviewable evidence package for an ELN, LIMS or records system
Checked 2026-09-01 Read the sheet

Work around the products

Four areas of engagement, each built on a product in the portfolio.

  • Vision system engineering
  • Model lifecycle
  • Edge and fleet deployment
  • Laboratory evidence software and validation support
Read what each one covers

What is actually true today

Every note below is quoted from the product's own repository README and dated. Two of the five are design work with nothing qualified in them, and they say so here.

Every product read from its repository on 2026-09-01.

Product Note, quoted from the repository
VisionAI “MVP security: session cookies, hashed passwords. Intended for a trusted LAN, not direct internet exposure.”
VisionML “ONNX export (placeholder graph today; real export = Phase 6)”
VisionEdgeOS “Status: design + scaffolding only. Nothing here is qualified, secure, immutable, fail-closed, or production-ready.”
VisionFleet “Every development-control response reports deployment_eligible=false. It provides no production approval, qualification, retention, evidence-disposition, or general remote-shell authority.”
Digital Labs “Until an approved artifact exists the app runs a deterministic preview provider and says so on the capture surface.”

Open the full ledger

Tell us what you need to inspect, count or prove, and we will tell you what it would take.

claudio@dantonoli.com

No form, no tracker. Mail goes straight to the person who built these products.

Write to VisionQ