It checks availability, books the table, takes the cancellation, and hands off the birthday party of fourteen to a manager, all in a normal conversation instead of a phone tree. No missed calls during the dinner rush, no answering-service script reading.
Most restaurant calls fall into a small set of patterns: check availability, book, cancel, ask a common question. The Scripted tier handles those with a fast, pre-built reservation flow that rebuilds every night from your actual availability, and quietly hands the rare unusual call to a full conversational fallback. The Live LLM tier runs every call through a live conversational agent from the first word, so nothing about the call needs to fit a pattern at all. Both tiers use the same voice, the same restaurant profile, and the same phone number setup.
Draft pricing, shown here for discussion. Final numbers depend on call volume and how many restaurants share the Scripted tier's infrastructure.
Vendor pricing pulled from public pricing pages, 2026-07-22. Effective per-minute rates include what each platform's own headline number leaves out (LLM, STT, TTS, telephony pass-through).
| Platform | Pricing model | Effective all-in rate |
|---|---|---|
| This product, Scripted tier | $199/mo flat | No per-minute cost for the reservation flows that make up most calls |
| Vapi | Usage platform fee + pass-through | ~$0.13–0.30/min in production |
| Bland AI | Bundled per-minute | $0.11–0.14/min, plus $299–499/mo at higher tiers |
| Retell AI | Componentized à la carte | ~$0.07–0.31/min blended |
| Synthflow | Enterprise only | Starting at $30,000/year |
| Coverage option | Typical monthly cost |
|---|---|
| In-house host covering phones during dinner shift | ~$2,000–2,500/mo in labor, before overtime or training turnover |
| Off-site human answering service | ~$300–700/mo for small-business call volumes |
| This product, either tier | $199-399/mo, answering calls around the clock, well beyond a single staffed shift |
Human-staffing figures are general industry estimates rather than vendor-published numbers. Refine with real local wage data before using in a signed proposal.
Scripted is the right starting point for most single-location restaurants. You can move to Live LLM later without redoing setup.
20-30 minutes to gather your hours, table layout and capacity, large-party policy, cancellation policy, and anything callers commonly ask about.
Forward your existing line, or port it over. Either way your listed number, website, and Google listing stay exactly as they are.
You call the agent before any customer does, working through a short list of scenarios. We fix anything that sounds off. You decide when it's ready.
The agent picks up on your number, following the hours and policies from setup. You can still answer yourself any time you want.
| Tier | Standard issue | Service-affecting outage |
|---|---|---|
| Scripted | Next business day | Same business day |
| Live LLM | Within 4 business hours | Within 2 hours |
| Enterprise | Within 1 hour (business hours) | Immediate, 24/7 |
A service-affecting outage means the agent isn't answering, or fails on the basic reservation flow. Every ticket gets a confirmation and a case number. If a call goes badly, tell us roughly when it happened; we pull the recording and transcript, confirm what happened, and report back exactly what we changed to fix it.
Direct write access into OpenTable, Resy, SevenRooms, and similar platforms is gated behind each vendor's own partner-approval process; none of them offers instant self-serve signup for booking-write access. We've mapped the real timelines for each provider (OpenTable is the most realistic near-term path, at roughly 3-4 weeks for partner review once an application is submitted) and we're treating this as a parallel business workstream that runs alongside the engineering build, without overstating it as already connected.
In the meantime, every reservation the agent takes is logged and confirmable the same call, and can be read into your existing system by your staff exactly as a phone-in reservation is today. We'll update this page the moment direct integration goes live for a given provider.