01
Same product, same releases.
Not a marketing build, not a trimmed preview. It is deployed by the same factory, from a tagged release and a clean checkout, and it reports its own version like any other deployment.
The product's own first deployment, running the same releases every other deployment runs.
Look before you talk to anyone.
Most private AI is sold from a slide deck and a screen share. The first deployment of this platform is our own, it is the same product a client gets, and it is open. Register with an email address and go through it yourself.
Our own deployment, not a sandbox we built for you.
THE DEMO · WHOSE DEPLOYMENT THIS IS
The platform's reference deployment is also the live demo. It runs from the same tagged releases, the same content policy enforcement and the same administration model as every other deployment, which is the only reason it is worth your ten minutes. It is stood up by the same factory, from a tag that exists in a repository, and it reports its own version like any other deployment does.
02
Open, on purpose.
An email address gets you in. Form your own view of the product first, then decide whether a conversation is worth booking.
03
Not your environment.
Your brand is not on it, your documents are not in it, your people are not in its directory, and none of your modules exist here. Treat everything you type as visible to us and to other visitors.
What is here, and what your own deployment adds.
WHAT YOU SEE · DEMO AND DEPLOYMENT
| Surface | In the demo | In your own deployment |
|---|---|---|
| The dashboard | ✓Ours, in our brand | ✓White-labelled at build time. Your name, not ours |
| Chat and the model selector | ✓Streaming, sessions, attachments, model and region choice | ✓Same, over the model set decided with you |
| Knowledge base toggle | △Over documents we loaded, shared with every visitor | ✓Over your documents, readable the same way by every user of your deployment |
| Usage view | ✓Your own tokens, quota gauge, by day | ✓Per person, plus usage by subject for administrators |
| In-product docs and the "your plan" page | ✓Our deployment's resolved controls and standing exceptions | ✓Yours, generated from the same typed object that builds your infrastructure |
| Administration | ✕Not open to a visitor | ✓Users, roles, grants, quotas, audited transcript reads |
| Modules | ✕Nothing admitted for you to see | ✓Admitted per deployment, on their own origins |
| Your data | ✕None of it, and please keep it that way | ✓At rest in a cloud account you own and are billed for |
The row marked with a triangle is the one worth reading twice: it describes the product, not the demo. Chat and knowledge, with its limits →
Five things worth trying first.
TEN MINUTES · A SUGGESTED ORDER
Do these in order
Ten minutes is enough to know whether this is serious.
- 01Switch models mid-conversation. Open the selector, read what is offered and what is marked unavailable, and change model on the next message. Switching a model is configuration here, not a migration.
- 02Turn the knowledge base on. Ask something the loaded documents answer, then ask something they cannot. Watch it decline rather than improvise, and read the citations it returns.
- 03Open your usage. Tokens by day, a quota gauge, document count. There is no currency anywhere in the product, and that is deliberate.
- 04Read "your plan". It renders that deployment's resolved controls and its standing exceptions, generated from the same typed object that builds the infrastructure.
- 05Try to find the marketing. There is none inside the product, because it is the product rather than a tour of one.
The one that tells you the most
Step four. Almost every AI product will show you a chat window. Very few will show you, inside the product, a generated statement of which controls are on, which are not, and what remains unclosed on that specific deployment. We show you, inside the product, what we have not yet closed.
Please do not upload anything sensitive
Everything loaded into a deployment's knowledge base is readable by every user of that deployment, and this one has other visitors in it. Use documents you would be comfortable handing to a stranger.
A demo can only prove the visible half.
WHAT IT PROVES · AND WHAT IT CANNOT
A
What it proves.
- The product exists and is finished enough to hand to a stranger
- Chat, model choice, the knowledge base and metering are real surfaces, not screenshots
- The controls page is generated rather than written, and it names its own exceptions
- Nobody had to schedule anything for you to see it
B
What it cannot.
- That it is in an account you own, which is the whole argument
- Your content policy, your residency choice, your posture tier
- Your roles, your grants, your quotas, your audit record
- Your modules, which are the part people use every day
It's not a black box. You can see how it's set up, and that's exactly what lets you answer for it.Why the demo is open at all
- 01Platform overviewWhat lands in your account, and who owns it afterwards.→
- 02Posture tiersBaseline, Standard, Strict, and what each one actually buys.→
- 03Pricing$2,500 assessment, deployment from $2,500, $500 a month, compute billed to you.→
- 04Platform or gatewayIf your people already use an assistant, the other product may be the right one.→
⎯⎯ Book the strategic assessment ⎯⎯
Your private AI, inside your control ·
Once you have looked, the assessment is the bounded step that turns what you saw into a deployment shape for your own account. You keep the deliverable either way.