Production duplex speech model for revenue calls
Details
- External ID
- 49026899
- Source
- HN
- Company
- —
- Product
- Production duplex speech model for revenue calls
- Website domain
- metavoice.io
- Launched
- July 23, 2026
- Cohort
- —
- Upvotes
- 14
- Upvotes percentile
- 0.6338112305854241
- Tags
- —
- Fetched at
- Sept. 7, 2026, 9:26 p.m.
- Updated at
- Sept. 7, 2026, 9:26 p.m.
Description
Hi, HN, we’re building Mia & Leo: AI personalities backed by purpose-built duplex speech models for revenue calls.Teams building real-world agents told us that more than 40% of callers hang up within the first 30 seconds. On a revenue call, that meant a lost booking, an unpaid balance, a cold lead, or a customer who does not come back.This happens because humans interrupt, correct dates, pause halfway through a sentence, or speak over someone in the background. The agent misses it, talks over them, or loses the thread.That’s because of how the current voice systems work. They sequentially listen, stop, think, and then speak. The conversation is actually a walkie-talkie exchange.So, we’re proud to introduce our first iteration of Mia & Leo who keep listening while they speak. They can handle interruptions, overlapping speech, corrections, and background voices without breaking the conversation.They are also built for production. Developers get control over agent behaviour, visibility into calls, failures they can debug, and costs that work at scale.Please try them out at metavoice.io. We'd love to hear any feedback!
Enrichment
- Theme
- ai agents for calls and meetings
- Vertical
- Sales
- Function
- Agent / copilot
- Audience
- B2B
- AI stance
- AI-native
- Project type
- Commercial product
- Normalized one-liner
- duplex speech model for revenue calls
- Manually corrected
- False
Could you build this?
No Building production-grade full-duplex speech models for low-latency live telephone conversations requires frontier multimodal generative audio ML research and heavy infrastructure.
What it would actually take: The architecture demands an end-to-end full-duplex neural audio model (or ultra-low latency cascaded pipeline with streaming VAD, turn-taking models, and streaming TTS) connected directly to SIP/WebRTC telephony backends. The core challenge is maintaining sub-300ms turn-around, continuous conversational interruption handling, and natural barge-in without audio artifacts. This requires millions of dollars in compute, specialized speech/audio ML researchers, and low-level WebRTC/telephony infrastructure engineers.
Discussion
2 comments analyzed.
Concerns raised: Lack of public benchmarks or performance metrics
Competitors
Other products that read as similar to this one — 134 launches clear the similarity bar, closest 8 shown.
Attention rank: #65 of 135 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 258 days after the earliest competitor.
- I built a voice AI that responds like a real woman · hn · 2026-03-25 · 6 upvotes · similarity 0.47
- Sparrow-2 · hn · 2026-08-27 · 8 upvotes · similarity 0.45
- Sparrow-2 · hn · 2026-09-08 · 11 upvotes · similarity 0.45
- Voice gender classifier for European voice AI (1MB, ONNX, 4ms) · hn · 2026-05-12 · 5 upvotes · similarity 0.44
- Τ³-Bench is out · hn · 2026-03-25 · 12 upvotes · similarity 0.43
- Vocal · ph · 2026-09-22 · 4 upvotes · similarity 0.42
- Speak Chinese: AI Voice Chat · ph · 2026-09-17 · 1 upvotes · similarity 0.42
- VoicePilot · ph · 2026-09-06 · 1 upvotes · similarity 0.41
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a agent / copilot tool for Agriculture yet.