#voice-ai articles

Pritesh Kumar23h ago · 6 min read

How to Stop Your Voice Agent Interrupting Callers

If your voice agent talks over callers, it's ending their turn too early, and the fix is end-of-turn detection that listens for more than silence, plus a longer wait before it replies to a turn that s

dograh.hashnode.dev

VoxRT1d ago · 14 min read

Building a fully offline voice assistant on Raspberry Pi 4

What we are building A voice assistant that runs entirely on a Raspberry Pi 4. No cloud, no API keys, no per-minute pricing. The user says "Hey Assistant", the device wakes up, and then handles whatev

voxrt.hashnode.dev

Neeraj Sharma1d ago · 4 min read

100 Clinics, One GPU Pool: A System Design for a $30K MRR AI Receptionist on Open Source

The pitch is easy: 100 small businesses × $300 a month = $30,000 MRR for an AI receptionist that answers every call. The system design is the part that decides whether you keep those 100 customers. Th

axionry.hashnode.dev

Apptycoons2d ago · 12 min read

AI receptionist call routing: from demo to booked appointment

Quick answer: In production, an AI receptionist works best as a front agent with one job: find out what the caller wants and hand the call to an agent built for that request. The agents behind it stay

apptycoons.hashnode.dev

Apptycoons3d ago · 12 min read

Why sub-second latency makes or breaks a voice agent

Quick answer: Aim for the agent to start speaking within about one second of the caller finishing, and faster where you can. People leave gaps of roughly 200 ms between turns, so longer silences feel

apptycoons.hashnode.dev

Apptycoons3d ago · 13 min read

Xtime API integration for AI voice agents: a build guide

Quick answer: Xtime has no self-serve API. As of October 2026, access runs through Cox Automotive’s partner program: an NDA, a partner agreement, fees, a business review and a certified demo, then an

apptycoons.hashnode.dev

Neeraj Sharma4d ago · 6 min read

200 Milliseconds: A System Design for Voice Agents Built From Many Small Models

On October 19, monday.com launched Harmony, a voice AI platform for sales calls. The interesting part is not the product. It is the architecture: instead of one general-purpose model, Harmony uses a s

axionry.hashnode.dev

shah Diya5d ago · 6 min read

The Journey of a Voice: From Your Microphone to an AI Agent

Everything a voice AI app does, from a phone call to a speech model, rests on a few simple ideas: sound is a wave, a computer stores it as numbers, and a network ships those numbers in small packets.

diyaa.hashnode.dev

VoxRTOct 23 · 10 min read

On-device streaming ASR in 2026: five approaches and when to pick what

Why this topic right now In 2023 there were two options. Take OpenAI Whisper and live with offline batch inference (3+ GB large-v3, not streaming), or wait for cloud ASR with its privacy, cost, and la

voxrt.hashnode.dev

Neeraj SharmaOct 21 · 7 min read

Voxtral 3 Prices a Voice Agent at 4 Cents a Minute. The Models Inside Cost 1.5

Mistral shipped Voxtral 3 this week with something new in its price table: a full voice agent API, priced per minute. With Mistral Small 4 doing the reasoning it costs 0.04 USD a minute. With any othe

axionry.hashnode.dev

Search Hashnode

Trending tags this week

How to Stop Your Voice Agent Interrupting Callers

Building a fully offline voice assistant on Raspberry Pi 4

100 Clinics, One GPU Pool: A System Design for a $30K MRR AI Receptionist on Open Source

AI receptionist call routing: from demo to booked appointment

Why sub-second latency makes or breaks a voice agent

Xtime API integration for AI voice agents: a build guide

200 Milliseconds: A System Design for Voice Agents Built From Many Small Models

The Journey of a Voice: From Your Microphone to an AI Agent

On-device streaming ASR in 2026: five approaches and when to pick what

Voxtral 3 Prices a Voice Agent at 4 Cents a Minute. The Models Inside Cost 1.5