Development simulation Fake compute, fake model replies, and simulated credits only. Nothing here creates a charge.
M1 local demo

AI, WHEN YOU NEED IT

A little help.
Or a lot of horsepower.

Choose the right workspace for your next task. See the server-approved spending cap before your session begins.

15 min4 hours
Model and session limits

Loading server-owned quote…

This is deterministic local inference. Proposed model limits require later performance testing. Startup and canceled startup cost C$0.00.

PAY FOR A SESSION

Pick the power.
Keep the spending visible.

Chat is the only M1 tier. The other proposed tiers remain visible so the product direction and price transparency are preserved.

One session, one clear rate

The accepted quote pins the CAD rate and cap on the server. Changing browser text or timers cannot change accounting.

No charge for warming up

The meter starts only when the fake model is ready and you explicitly enter. Failed or canceled startup costs nothing.

End cleanly

End is idempotent, settles the simulated charge once, and releases the fake compute allocation.

BEHIND THE SESSION

Ready when available. Started when needed.

01

Accept a quote

The backend resolves the profile, proposed rate, 15-minute minimum and spending cap.

02

Wait for readiness

The browser polls server state; ready means the fake provider reports its model prepared.

03

Enter and wind down

Explicit entry starts accounting. End or cap settlement happens once and releases fake compute.