Gemini for Home debuts this October

Author

Ryan Howard
August 28, 2025

🔊 Soundcheck

  • Gemini for Home debuts this October

  • Vox AI’s $8.7M push into fast-food voice tech

  • Loman AI scores $3.5M seed round

  • Deepgram voice tools join Workers AI

Read time: 4 minutes

🔥 Hot Mic

Big moves, deep dives, and standout stories.

Google’s Gemini for Home rolls out in early access this October, replacing Assistant with smarter AI help.

Google has rolled out “Gemini for Home,” a new AI voice assistant powered by its most advanced Gemini models. It’s poised to replace Google Assistant on existing smart speakers and displays and promises much smarter, more conversational home control.

The assistant handles nuanced, multi-step instructions fluently—from media playback to smart home automation—and even lets you say things like “turn off the lights everywhere except my bedroom.”

Gemini Live enhances the experience with back-and-forth dialogue—no more repeating “Hey Google” between questions. You can troubleshoot appliances, brainstorm ideas, cook interactively, or even craft bedtime stories in real time.

Google plans to launch this in early access in October 2025, with both free and paid tiers available eventually. It’s a significant push toward more capable and natural AI throughout the home.

Key Points:
• Gemini for Home replaces Google Assistant on smart speakers and displays.
• Handles complex, natural-language requests across smart home devices.
• Gemini Live enables ongoing conversational interaction without repeat prompts.
• Early access begins October 2025; will offer free and paid tiers.

Takeaway: Gemini for Home marks a meaningful leap in home voice AI—making smart home control feel conversational, capable, and context-aware—ushering in a new era beyond rigid commands.

Vox AI raised $8.7M in seed funding led by Headline to expand its autonomous drive-thru voice platform worldwide.

Vox AI, a conversational voice-AI startup for quick-service restaurants, secured $8.7 million in seed funding, led by Headline, with support from True, Simon Capital, and returning investor Souschef Ventures. This capital moves their total to $10 million and powers their global roll-out—including a new San Francisco office. Their platform manages drive-thru and mobile voice orders in over 90 languages, offers employee support tools, and integrates into existing tech stacks. Early deployments already report up to 17× ROI, shorter queues, better upselling, and happier customers while freeing staff for higher-value work.

Key Points:
• Raised $8.7M in seed funding to accelerate global expansion.
• Lead investor: Headline; additional: True, Simon Capital, Souschef Ventures.
• Funds bring total raised to about $10M.
• Platform supports drive-thru/mobile orders, staff assist in 90+ languages.

Takeaway: Vox AI’s new funding round positions voice as the next major interface in fast food—boosting efficiency, revenue, and staff effectiveness across QSRs globally.

Loman AI raised $3.5 M to scale its voice-AI phone agent improving orders, upsells, and labor.

Austin-based Loman AI has captured $3.5 million in a seed round led by Next Coast Ventures, joined by TenOneTen Ventures and Antler. The financing will accelerate product development and support growing demand for Loman’s voice-AI phone agent across restaurants. The platform, handling millions of calls and tens of millions in orders since mid-2024, helps operators recapture missed calls, boost sales, and cut labor costs. Results so far include up to 22% more revenue from recovered calls and upsells, and labor savings of up to 17%, especially in high-volume segments like pizza.

Key Points:
• Seed round of $3.5 million led by Next Coast Ventures, joined by TenOneTen and Antler.
• Supports product development and team expansion amid accelerating demand.
• Platform processed millions of calls and tens of millions in order volume since 2024.
• Operators report up to 22% revenue increase and 17% labor cost reduction.

Takeaway: Loman AI converts the phone—from a liability to a revenue driver—by answering every call, integrating smoothly with existing systems, and delivering measurable ROI for restaurants.

Cloudflare's Workers AI now includes Deepgram’s text-to-speech (Aura-1) and speech-to-text (Nova-3) models, enabling real-time voice interactions with low latency and expressive output. These models run directly on Cloudflare’s global GPU-powered edge infrastructure.

Developers building voice agents can leverage features like WebSocket support for streaming audio and turn detection for detecting when someone stops speaking—making dynamic, responsive voice AI easier to build.

Key Points:
• Workers AI now hosts Deepgram’s Aura-1 (TTS) and Nova-3 (STT) models.
• Deepgram’s models provide fast, expressive voice generation and accurate transcription.
• WebSocket support enables bi-directional, real-time audio streaming.
• Turn detection model helps determine when speech ends for smoother voice agents.

Takeaway: By embedding Deepgram’s voice AI directly into Workers AI, Cloudflare sharpens its edge as a low-latency, serverless platform for real-time voice agent development.

🎙️ Mic Drop

What else is making noise in voice AI.

  • SuperDial raises significant funding to further R&D and scale its healthcare-focused voice AI platform. mobihealthnews.com

  • Detroit-style pizza chain uses Kea’s conversational AI to automate all phone orders, showing at-scale restaurant voice AI deployment. restauranttechnologynews.com

  • Kea powers customer service and order consistency as VIA 313 expands to new locations using AI. restaurantnews.com

  • Analysis of SoundHound AI’s voice technology and business strategy to maintain long-term competitive advantages. ainvest.com

  • Google develops AI agents to streamline customer support with automated voice and chatbot experiences. cxtoday.com

  • Commerce.AI to present on production agentic voice AI saving millions for enterprise customer service. accessnewswire.com

  • Step-by-step Python guide for building a custom AI voice assistant, aimed at developers and technical teams. geeky-gadgets.com

  • Ostro debuts AI that turns static marketing emails into voice-enabled, compliant conversational experiences for pharma marketers. morningstar.com

  • Market overview: More startups and investments are pouring into restaurant voice AI solutions. restaurantbusinessonline.com

  • Comprehensive guide to Apple's next-gen voice assistant powered by large language models. macrumors.com

  • Tutorial on using CapCut’s voice generator to create engaging game recaps without recording your own voice. clnsmedia.com

  • Investment analysis on SoundHound AI, highlighting its financial positioning and future potential in voice AI. seekingalpha.com

  • Longstanding customer service provider Moneypenny deploys new AI agent aimed at human-like call handling. businesscloud.co.uk