Voice Kit · TypeScript voice-agent kit
AI Voice Receptionist Kit
Answer every call for your clients — on your own keys, at your own margin.
27% of calls to home-services businesses go unanswered, and under 3% of callers sent to voicemail leave a message — that's booked jobs walking away. Agencies already invoice $250–500 a month for AI receptionists, mostly built on platforms that take $0.07–0.31 a minute. This kit is the per-client deployment system on your own keys instead — roughly $0.04–0.12 a minute all-in: the Twilio ConversationRelay + realtime-LLM pipeline, per-client config files (no code forks), booking and lead intake, after-hours mode, and the part free templates skip entirely — consent-safe missed-call text-back defaults and the A2P 10DLC onboarding checklist every client deployment needs.
Built for: Developers and small agencies deploying AI phone receptionists for local service businesses
- Twilio ConversationRelay + realtime LLM pipeline — bring your own keys
- Per-client configuration files — services, hours, pricing, escalation — no code forks
- Booking and lead intake that actually books
- After-hours and overflow answering modes
- Consent-safe missed-call text-back: informational first text, quiet hours, opt-out
- A2P 10DLC per-client onboarding checklist
- Margin worksheet: your keys vs platform per-minute pricing
- 78automated tests green at packaging
- 1runtime dependency — ws, nothing else
- 2example client configs, ready to copy
- 12 moof updates against moving APIs & rules
Pick your license
Solo
$79
One developer, unlimited deployments.
- The full kit
- Voice pipeline + per-client config
- Compliance defaults + checklist
- 12 months of updates
One developer's own projects.
Team
$149
Your whole team deploys with it.
- Everything in Solo
- Team-wide license
- Priority email support
One team, internal and client deployments.
Agency
$349
About one client-month. Unlimited clients.
- Everything in Team
- Unlimited client receptionists
- Source design files where provided
Unlimited client deployments.
14-day money-back guarantee — if the kit doesn't fit your stack, reply to your receipt within 14 days and we refund it in full. Keep nothing, owe nothing.
More from the store
Agent Eval Harness Kit
The eval layer free runners leave as an exercise — task sets, rubrics & a CI gate that says no.
Signal Foundry — AI Product Studio Kit
The 40 hours before your first prompt — research, architecture, and an app that refuses to make things up.
Relay — Autonomous SDR
The AI SDR that can't make things up.
Questions
What exactly do I get?
An instant download: the TypeScript kit — the Twilio ConversationRelay + realtime-LLM voice pipeline running on your own API keys (Anthropic and OpenAI adapters included), per-client config files with a strict validator and two ready examples, booking and lead intake that lands in your webhook plus a local JSONL fallback, after-hours and overflow modes, the consent-safe text-back module, the A2P 10DLC onboarding checklist, a margin worksheet, real docs, and 12 months of updates. npm test runs typecheck plus 78 tests — green at packaging on a clean install.
What do I still need to bring?
Your own Twilio account and LLM API keys (that's the point — the margin stays with you), a public URL for the webhooks (cloudflared or ngrok locally, any host in production), and each client's A2P 10DLC registration, which the checklist walks step by step. Honest note: packaging QA is an offline test battery plus schema verification against Twilio's docs — place your first live test call before you put a client on it, and the quickstart shows exactly how.
Is the missed-call text-back TCPA compliant?
We ship consent-safe defaults — an informational-only first text replying to a call the customer made, quiet-hours enforcement, and opt-out handling — plus a per-client checklist including A2P 10DLC registration. That raises the floor; it is not a legal guarantee, and the rules moved as recently as February 2026. Final compliance review for each deployment is yours.
Was this scoped by AI?
Yes — our commercial engine researched it: the missed-call data, platform pricing, the reseller economics, free alternatives, adverse evidence included, all source-checked. Full honesty: the engine scored this concept below our own demand gate — strong demand and growth scores, but one-time kits capture little of this market's recurring value — and the owner shipped it anyway as a bet. What's not a bet is the code: 78 tests green on a clean-room install at packaging, enforcement behaviors actually tested, and the 14-day refund applies.