Every team that adds calling to their product rebuilds the same backend: token services, webhook routers, call control, and a dialer UI. Alloqui turns all of that into one component and one project key.
Adding voice calling to an app sounds like a feature. In practice it is a project.
Pick a provider — Twilio, Plivo, Amazon Chime — and the SDK is the easy part. Before your first call connects, you also need:
None of this is your product. All of it stands between you and shipping.
We looked hard before building anything.
Per-minute AI platforms like Vapi and Retell are great if you want an AI agent answering calls — but you pay a per-minute markup on every call forever, and you are locked into their stack.
Open-source SIP and voice libraries give you a transport layer and stop there. You still build the backend, the token flow, and the entire UI.
Provider SDKs themselves are solid but provider-shaped. Building on @twilio/voice-sdk directly means Twilio concepts leak all over your codebase, and switching providers later is a rewrite.
The gap: nobody offered a drop-in dialer where you bring your own provider keys and pay a flat fee — not per minute.
Alloqui is the "Firebase for Voice" model:
al_live_...import { Dialer } from "@alloqui/dialer";
<Dialer projectKey="al_live_abc123" />
That's it. Tokens, webhooks, call control, events, and the UI are handled.
Alloqui only handles signaling — lightweight JSON that sets calls up and tears them down. The actual audio flows directly between the browser and your provider over WebRTC.
This matters for three reasons:
Under the hood, every provider implements the same interface — on the backend (token generation, call control) and in the browser (a VoiceEngine abstraction that lazy-loads only the SDK your project needs). Your project key resolves to the right provider at runtime.
Today that means Twilio and Plivo. The abstraction is the point: when we add a provider, you switch by changing credentials in the dashboard — not your code.
The roadmap is public and simple: inbound calling and recording on Pro, more providers, and AI transcription and summaries with your own STT/LLM keys — again, no per-minute markup on intelligence either.
If you have rebuilt telephony plumbing more than once, start building — the free tier is genuinely free.