What does a two-minute voice clone actually sound like?

Hi ,

xAI just made voice cloning simple enough for any developer to ship.

Record one minute of natural speech. Get a production-ready voice model back in under two minutes.

Use it across text-to-speech, voice agents, and real-time conversational AI. No extra charge.

Today's prompt writes podcast ad scripts that sound like the host actually means it. Tool Tuesday covers Google's new Gemini Webhooks for long-running AI jobs. Then what xAI's custom voice feature means for developers and creators.

🔥 Prompt of the Day 🔥

Podcast Ad Read Script Generator: Use ChatGPT or Claude

Create one host-read sponsorship template.

"Act as a podcast advertising specialist. Create one host-read ad script framework for [PRODUCT] that sounds authentic, not scripted.

Essential Details:

• Product Type: [WHAT YOU'RE PROMOTING]

• Podcast Genre: [SHOW CATEGORY]

• Host Voice: [PERSONALITY MATCH]

• Ad Length: [30/60/90 SECONDS]

• Promo Code: [TRACKING MECHANISM]

• Desired Tone: [CASUAL/EDUCATIONAL]

Create one ad script template including:

• Natural transition openers (5)

• Personal story placeholder

• Problem-solution bridge

• Benefit highlight (non-salesy)

• Promo code integration

• Natural endorsement close

Sound like a friend, not an advertiser."

Variables:

PRODUCT: What you're promoting

SHOW CATEGORY: The genre of podcast running the ad

PERSONALITY MATCH: How the host typically sounds

AD LENGTH: 30, 60, or 90 seconds

PROMO CODE: Your tracking mechanism

DESIRED TONE: Casual, educational, or somewhere between

Why This Works:

Listeners skip scripted ads. They stay for genuine recommendations. AI builds the framework that gives hosts the structure they need without removing the personality that makes it land. Natural openers. Real story placeholders. A close that sounds like advice not a pitch. The best podcast ads don't sound like ads at all.

🤖 Tool Tuesday 🤖

Google Launches Gemini API Webhooks — No More Polling for Long-Running Jobs

Building agentic apps with Gemini just got significantly less painful.

Gemini API Webhooks are live today. Available to all developers. No extra cost.

Push-based notifications that fire the instant a long-running job completes. No more repeatedly calling GET to check if your task is done.

The Problem It Solves

Agentic workflows — Deep Research, long video generation, batch processing thousands of prompts — take minutes or hours to complete.

Until today developers had to poll continuously. Call the API. Wait. Call again. Wasteful, inefficient, and adds latency to every long-running job.

Webhooks flip that entirely. Gemini pushes a real-time HTTP POST to your server the moment a task finishes. Your code waits. Gemini calls you.

How It Works

Configure webhooks globally at the project level or override dynamically per request for specific jobs.

Every request signed using webhook-signature, webhook-id, and webhook-timestamp headers. Idempotency guaranteed. Replay attacks prevented.

At-least-once delivery with automatic retries for up to 24 hours. If your server is down when the job completes, Gemini keeps trying until it gets through.

Works with Batch API jobs, Deep Research, long video generation, and any other long-running Gemini operation.

Why This Matters

Polling is a workaround. Webhooks are infrastructure.

For developers: Cleaner code. Lower latency. No wasted API calls checking job status every few seconds.

For production applications: Reliable delivery with retries means jobs don't silently fail when your server has a hiccup.

For the Gemini ecosystem: The difference between a model people experiment with and one they build on seriously is developer experience. This is that difference.

Full documentation and an end-to-end cookbook are live on the Gemini developer site now.

Long-running jobs just got a lot easier to build around.

Did You Know?

AI systems can now reconstruct a rough image of what someone is looking at by analysing their brain activity through a non-invasive scan — achieving results that would have been dismissed as science fiction a decade ago, though the images remain blurry and the technique is far from practical use.

🗞️ Breaking AI News 🗞️

xAI Launches Custom Voice Cloning — Your Voice Ready in Under Two Minutes

xAI just made voice cloning accessible to any developer.

Custom Voices and a new Voice Library launched inside the xAI console this week.

Record about a minute of natural speech. Production-ready voice model delivered in under two minutes. No extra charge on top of existing TTS or Voice Agent API pricing.

How It Works

Head to the xAI console. Record roughly a minute of natural speech.

Two-stage verification runs automatically.

Stage one — read a passphrase aloud. The system transcribes and matches it in real time. Confirms consent and presence.

Stage two — speaker embeddings from the passphrase and full recording are compared to confirm they belong to the same person.

You cannot clone from a pre-existing recording. You cannot clone someone else's voice. Both checks happen before the voice model is created.

Once verified your custom voice is ready instantly. Pass the voice ID to any TTS endpoint or use it with the Voice Agent API for real-time conversational agents.

What Your Custom Voice Can Do

Works everywhere xAI's built-in voices work.

Speech tags for pacing and emphasis. Multilingual output across 28 languages. REST and WebSocket streaming for real-time use.

Build a customer service agent that sounds like your brand. Create training content narrated in your own voice without re-recording everything. Power a sales assistant that speaks consistently across every conversation.

The Voice Library

New console page where your team browses, previews, and manages every voice available to your account.

Custom voices sit alongside 80-plus built-in voices across 28 languages. Preview any voice across different scenarios before choosing one for your application.

Why This Matters

Voice is the last major friction point in building AI-powered audio experiences.

Until now you either used a generic AI voice that felt impersonal, paid for expensive custom voice work, or recorded everything manually.

Under two minutes from recording to production-ready voice model at no extra cost changes that calculation entirely.

For content creators: Your voice. Your brand. Consistent across every piece of audio content you produce without re-recording.

For developers: A voice that matches your product's personality without the studio cost or timeline.

For businesses: Customer-facing AI agents that sound like your company instead of a generic assistant.

What This Means

Voice cloning used to mean expensive studios, weeks of work, and proprietary contracts.

xAI just made it a two-minute task anyone can do from their browser.

Content creators get consistent audio across everything they produce without sitting in front of a microphone every time.

Developers get a voice that actually matches their product instead of a generic assistant that sounds like every other AI.

Businesses get customer-facing agents that sound like their brand from day one.

And nobody pays extra for it.

Your voice just became an asset you can deploy anywhere.

Over to You...

xAI just made voice cloning free and instant.

Does that change anything for you?

Reply and share your take.

To staying ahead,

Jeff J. Hunter 
Founder, AI Persona Method | TheTip.ai

P.S. Want to turn AI Agents into a consulting offer? Book your AI Certified Consultant strategy 👉 here.

NEW: “AI Money Group” to Learn to Make Money with AI