Voice layer
Real-time speech in and out via Fire It Voice, with per-location configuration for pacing, disclosure, and transfer targets.
Fire It's ordering flow separates the voice from the truth. The model captures intent. The server enforces the menu, calculates the price, and decides when 'confirmed' is honest.
Real-time speech in and out via Fire It Voice, with per-location configuration for pacing, disclosure, and transfer targets.
Items, modifiers, sizes, quantities, and placement (whole / left / right) captured as fields — not free-text notes.
The server computes totals against your live menu. Neither the model nor the browser is trusted to price.
The agent uses 'let me repeat back' during, and only says 'you're all set' after the authoritative order state is recorded.
Certified POS integrations receive orders directly. When integration isn't available, orders land in the Fire It internal queue.
Every call gets a transcript, structured events, and a plain-English summary in the dashboard for QA and support.
The agent transcribes and understands the caller's intent in real time.
Menu, modifier caps, sold-outs, and required groups enforced against your live configuration.
Totals computed authoritatively server-side, then read back to the caller.
Order state moves through a finite state machine; only then does the caller hear 'you're all set'.