Glassware Labs
Glassware Labs · San Bernardino, California

Software that makes courses accessible. Sensors that hear what cameras miss.

Glassware Labs makes Glassability, the accessibility repair service for Canvas courses, and through Glassware Labs Engineering designs passive sensing systems that detect and locate aircraft without transmitting anything.

glass·ware n. The instruments of a laboratory: the vessels work is done in, made to be seen through. Also, what this lab makes.

Accessibility

Glassability™

An app inside Canvas. It reads a course the way assistive technology reads it, pages, documents, images, video captions, and measures what would fail a student who cannot see, hear or click the way the author does. Then it does the other half of the job the scanners leave behind: it drafts the repair, shows the before and after, and publishes only what the instructor approves.

What the college gets is not a promise but a receipt, in the numbers of the scanners it already runs. Built by a professor who teaches with it.

Glassability at accessibility.glass
715 → 44−94% findings
Scanner findings, before and after
Two pilot courses, measured by a third-party scanner's own re-scan, Fall 2026.
1,016,978
Caption words verified
Word by word, with human review where it counts.

Engineering

Glassware Labs Engineering

Passive sensing systems: arrays of microphones and radio receivers that detect aircraft by their sound signature and their radio emissions, find their direction, and hand the track to whatever the user already runs. Nothing in the kit transmits. Built for vehicles and fixed sites, developed in California.

About the engineering work
Contact
engineering@glasswarelabs.com

How we use AI

In Glassability, the accessibility gap persisted because the work was manual and nobody was staffed to do it. AI changed that arithmetic. We use it for the labor, and we keep people for the judgment. Three things are true of every model we run, and will stay true.

It works on course content, never on students. Pages, files, images, captions. Glassability never requests student submissions, grades or the people in a course, so no model can be shown them.
It drafts. A person decides. Every proposed change waits for the instructor's press, item by item. Every change is logged with a copy of the content before and after, and can be reversed.
The best AI, nothing to learn from. Course content is never used to train any model, and no student data reaches one. We use the strongest model for each task: the frontier models (Google Gemini, OpenAI, Anthropic) where they do the better work, under data processing agreements, in the United States, with no training; models on infrastructure we operate where those do as well. The point is the best work a student can get, with no new burden on faculty.

The Department of Justice has said fully automated remediation is not yet good enough to trust on its own. We agree; that is why the approve button exists. What Glassability keeps, and what the models see.

Who

Glassware Labs is the studio of Benjamin Wilbur, an electrical engineer and physician who teaches Human Services as an adjunct instructor at San Bernardino Valley College. Like most instructors at most community colleges, he had no instructional designer, no captioning service and no accessibility staff, only courses full of lecture video, slide decks and readings, and scanner reports listing hundreds of findings that were nobody's job to fix.

Glassability was built to be that missing job. It runs on his own courses first, and every number on this page came from them.

The name is a promise about how the instruments behave. Laboratory glassware holds the work without changing it and lets everyone see what is going on inside. Museum glass does the same for a painting. Glassability does the same for a course: unaltered except where the instructor said yes, and readable by every student who shows up.

The engineering work holds to the same habit: measure first, write it down, ship the piece that is proven.

Contact

Pilots, questions, and the people who run Canvas at a college: write to us with the contact form. Free pilot terms, the privacy page and the accessibility statement for Glassability are on its own site. For the engineering work, write to engineering@glasswarelabs.com.