Run SpeedPlus360 with zero AI — every module works without it. When you switch a feature on, you pick the model behind it: local models on your own private server, or cloud models from Anthropic, OpenAI or Google. Per feature, cost shown first, change anytime.
Per-feature cost shown before you enable. Change models anytime.
Receptionist, analytics, idea engine, reordering, assistants, voice — each is its own independent on/off switch in your admin portal.
Every module's core workflow runs with its AI features off. A kopitiam that wants a straight POS gets a straight POS — no dark patterns, no forced AI.
Run open-weight models on your own private server — your data never leaves the machine. Maximum privacy for the features you choose.
Anthropic Claude, OpenAI and Google Gemini available where you want the strongest models — connected through your platform's model gateway.
A local model for employee SOP answers, a frontier model for the voice agent — assign a model to each feature independently and mix freely.
Every feature shows what your model choice costs before you switch it on — you trade quality against price with the numbers in front of you.
AI usage is metered per feature per model, so the bill explains itself — you see exactly which feature spent what, in RM, every month.
Per-feature budgets put a ceiling on AI spend — a busy month raises questions, never a surprise invoice.
Swap the model behind any feature with instant rollback — upgrade when a better model arrives, step back the moment you prefer the old one.
All model calls flow through the model gateway on your own server, which holds your per-feature routing — one place to see and control every AI call.
Your restaurant runs on its own dedicated server with its own AI memory — no shared vector store, no other tenant's data anywhere near yours.
Start with zero AI and add one capability at a time as trust and value prove out — the dial goes from zero to full, and it's yours to turn.
Switch everything off and SpeedPlus360 is still a complete restaurant system — POS, reservations, scheduling with keyword WhatsApp commands, payroll, accounting, e-invoice. Nothing degrades, nothing nags. AI is an option you own, not a dependency you inherit.
Ready to try the AI receptionist? Open its switch, and the panel shows the model options side by side — a local model on your own server at hardware-only cost, or Anthropic, OpenAI and Google with estimated monthly cost for your volume. Pick one, set a budget cap, switch it on. The choice is informed, reversible and yours.
Prefer to keep employee SOP answers entirely on your own machine? Reassign that feature to a local model — data never leaves your server. A new frontier model just launched? Point the voice agent at it, and roll back instantly if you liked the old one better. Model choice is a setting, not a migration.
Swaps apply instantly, with one-tap rollback.
Be the first — join the waitlist and get onboarded first, on your own private server.