QCall.ai — Production AI Voicebot Platform
A full-stack AI voicebot with ~100ms STT, 50+ languages, Whisper Large v3 Turbo as LID, natural interruption handling, and a campaign orchestrator — built for QCall.ai to run automated sales and support calls at scale.
What QCall.ai was up against
QCall.ai needed a production voicebot capable of holding natural, interruptible conversations across languages with sub-second latency — something the open-source stacks didn't solve cleanly out of the box. Standard telephony providers offered only the plumbing, not the AI orchestration. The product had to handle concurrent calls, campaign scheduling, live transfer to humans, and rich analytics from day one.
What we built
We built the full voice agent stack — streaming STT with ~100ms latency supporting 50+ languages via Whisper Large v3 Turbo LID, LLM turn-taking with a custom interruption model, our open-source multilingual TTS (<300ms time-to-first-audio), barge-in handling, and a campaign orchestrator that drives outbound dialing. The dashboard surfaces per-call transcripts, sentiment, containment rate, and agent performance. Carriers are integrated via SIP trunks with automatic failover. Runs on a mixed Node.js / Python stack with Temporal for durable call-state machines.
What shipped
Want something similar?
Other work we’ve shipped
Telluswhatyouwanttoautomate.We'llreplyinonebusinessday.
Describe the problem, the constraint, the deadline. We'll send back a scoped plan and a senior engineer to kick it off — no sales theater.