DEV Community

#voiceai

Posts

đź‘‹ Sign in for the ability to sort posts by relevant, latest, or top.
Build a Low-Latency Voice Agent in Python with Pipecat

Build a Low-Latency Voice Agent in Python with Pipecat

1
Comments 1
9 min read
Why 500ms Latency Kills Voice AI: Inside Felona Voice's Sub-10ms Neural Routing Engine

Why 500ms Latency Kills Voice AI: Inside Felona Voice's Sub-10ms Neural Routing Engine

5
Comments
5 min read
Echo-cancellation warm-up eats a short first answer on phone calls

Echo-cancellation warm-up eats a short first answer on phone calls

Comments
6 min read
200ms Humans vs 700ms Voice Agents: What ACL IWSDS 2025 Says About the Turn-Taking Gap

200ms Humans vs 700ms Voice Agents: What ACL IWSDS 2025 Says About the Turn-Taking Gap

Comments
7 min read
Voice Agents in Microsoft Foundry: Inside the Realtime Speech-to-Speech Architecture

Voice Agents in Microsoft Foundry: Inside the Realtime Speech-to-Speech Architecture

Comments
19 min read
I Built an AI Agent That Talks Like Me. Here's What Actually Broke.

I Built an AI Agent That Talks Like Me. Here's What Actually Broke.

1
Comments
4 min read
I Built an AI Voice Sales Agent — Here’s the Architecture Behind It

I Built an AI Voice Sales Agent — Here’s the Architecture Behind It

1
Comments
8 min read
Nielsen's 3 UX Cliffs Mapped to Voice AI: 100ms Feels Instant, 300ms Alive, 800ms Dead

Nielsen's 3 UX Cliffs Mapped to Voice AI: 100ms Feels Instant, 300ms Alive, 800ms Dead

Comments 1
6 min read
How To Build Voice AI With Adaptive Latency For Poor Network Connectivity

How To Build Voice AI With Adaptive Latency For Poor Network Connectivity

Comments
1 min read
Three real sessions on a CPU voice agent API

Three real sessions on a CPU voice agent API

Comments
1 min read
Voice Pipeline Economics: Double-Billing, a Backwards Ladder, and a Lexicon That Never Reached the Voice

Voice Pipeline Economics: Double-Billing, a Backwards Ladder, and a Lexicon That Never Reached the Voice

Comments
5 min read
A Brand's Voice Becomes Persistent: Avatar Profiles Instead of Per-Run Re-Supply

A Brand's Voice Becomes Persistent: Avatar Profiles Instead of Per-Run Re-Supply

Comments
6 min read
From J/L-Cuts to One Call: Consolidating a Video Pipeline's Voiceover Path

From J/L-Cuts to One Call: Consolidating a Video Pipeline's Voiceover Path

Comments
8 min read
N-Person Cast: Teaching a One-Presenter Video Pipeline to Handle a Cast

N-Person Cast: Teaching a One-Presenter Video Pipeline to Handle a Cast

1
Comments
10 min read
Voice‑driven murder mystery: interview AI suspects with your voice

Voice‑driven murder mystery: interview AI suspects with your voice

Comments
2 min read
đź‘‹ Sign in for the ability to sort posts by relevant, latest, or top.