Your course's AI stack
Tuned to how you teach.
A writing seminar and a robotics lab shouldn't get the same AI model. Customize a base model with specific prompts and information so that each class gets a set of AI models customized for its work.
Customize the AI stack your class learns with.
Glassroom gives your course its own transparent and customizable AI stack — you design the base models, customize the information, and set the limits, so it works with, not against your teaching. The record of every exchange is visible and helps to surface insights the whole class can learn from.
Request a demo
Try it out with your class in mind.
Share your email and we'll be in touch with an invite to a guided Glassroom demo.
m.okafor · session 12today 14:32
Student
Why does my loop run one extra time and crash at the end?
Model
Classic off-by-one — your loop goes one index past the last item, so it runs out of bounds on the final pass. Loop to len(items) - 1, or iterate the list directly, and print the index each pass to watch where it overshoots…
cs-101 · class ledger24 students
| Student | Sessions | Exchanges | Last active |
|---|---|---|---|
| a.gupta | 11 | 198 | 11:05 |
| m.okafor | 12 | 214 | 14:32 |
| j.tran | 9 | 167 | 09:41 |
| s.whitfield | 7 | 122 | yesterday |
One record · two lensesRead one student's session end to end, or step back to the explore ledger of the whole class.
One gateway
One stack, one place for the class.
Students chat in a clean web interface, or point their own coding tools at an OpenAI compatible endpoint. You manage who has access and every request runs through a single gateway you can see — no personal accounts, no scattered logins, no API keys to hand out.
The record
Insights the whole room can use.
Each prompt and response is tied to a student and a conversation, and is visible to you. See how the class is using the model and what it surfaces about their learning. Use Glassroom's analytics to build custom queries to understand what students are understanding well, where they are struggling, and then bring those insights back to inform your teaching and better serve your students.
Your hardware
Runs on your hardware, on your terms.
Host Glassroom on your own server and AI inference resources and student data never reaches an outside vendor. No server or GPU on hand? Set up an account on our server and point it at a hosted AI model of your choice instead. Source-available end to end so you can see what's going on under the hood and customize it to your own environment as needed.
Get set up in an afternoon
STEP 1
Stand it up
Set up our software stack with a GPU server — or point it at a hosted model. About 20 minutes from a bare machine.
STEP 2
Create a class
Each class gets its own customized AI chat interface and private record — its own models, instructions, and roster, sealed off from every other class.
STEP 3
Add your roster
Students sign in with their school email. No accounts to buy, no keys to distribute, nothing for them to configure.
STEP 4
Source insights
Per-student sessions, whole-class patterns, usage counts — in a portal made for instructors, not infrastructure teams.
See the actual dashboard.
The screenshot below shows an instructor's view of one class. You can see who is using the AI tools and how, which models they're using, and see all the chats underneath.
Students know. That's the point.
Glassroom works because everyone can see through the glass. Students are told what's kept and who reads it — the record is a shared lab notebook, not a camera in the corner. Used well, it changes the conversation: instead of policing AI use you can finally teach about it and with it, with the actual transcripts on the table.
Built on vLLM, LiteLLM, and Open WebUI — proven open source components, assembled into one stack an instructor can run alone.