ElevenLabs can now clone your voice in 10 seconds 🎙️

Good morning. ElevenLabs dropped Eleven v4, a new voice model that clones you from 10 seconds of audio and speaks 90+ languages. A stunning AI landscape account is pulling millions of likes on calm, dreamy scenery set to soft music. And OpenAI shipped two cheaper GPT-6 models, Sol and Luna, for everyday work. Would you let an AI clone of your voice narrate all your videos? Hit reply and let me know! Let's get into it.

ElevenLabs dropped Eleven v4, its new text-to-speech model and the most expressive one it's made, plus a faster v4 Turbo for real-time stuff. It's live now in the app and API, and you can try it on a free account.

The headline feature is the cloning. It now needs only about 10 seconds of audio to clone a voice, and that clone can speak over 90 languages while still sounding like you, accent and all. The delivery is also way more human, it reads the emotion from your script on its own, and you can drop inline tags like [laughs], [whispers], or even [said angrily in French accent] to direct specific moments, plus background cues like [light rain]. It landed at number one on the main TTS leaderboard.

For creators this is a big one. Voiceover makes or breaks faceless content, and v4 is about as close to a real human read as AI has gotten, in basically any language you want. Clone your own voice once and you've got a narrator that never gets tired. If you do narration, explainers, or story content, it's worth testing this week.

OpenAI also rolled out GPT-6 Sol and GPT-6 Luna, two cheaper models in the GPT-6 family that sit below the flagship Astra. Sol is the everyday workhorse, Luna is built for high-volume tasks, and both come in around half the price of the models they replace.

They're trained with the same methods as Astra, so you get a good chunk of that quality for less, with Luna in particular being a surprisingly capable cheap option. They're aimed at work, scripting, and agent-style tasks, rolling out in ChatGPT Work, Codex, and the API, with Luna also available to users in the desktop app.

Same as the other model drops lately, this is a work and creative tool. If you run any kind of automated pipeline or lean on ChatGPT for the writing side, cheaper models with more quality is a quiet win. Just good to know the cheap tier got better.

You've definitely scrolled past these: impossibly beautiful landscape clips, a stone cottage in an endless sunflower field, misty mountains, golden-hour coastlines, set to soft, calming music with a short dreamy caption. They're AI, and they're pulling absurd numbers, one account's sunflower-field clip is past 5 million likes.

There's no hook to decode, no punchline, just a beautiful place and a soft sound that makes people stop scrolling and breathe for a second. That "I wish I was there" feeling is incredibly saveable and shareable, which is exactly what pushes a clip into millions of views, and the calm trending audio does half the work for you.

Here's the playbook. Generate a batch of dreamy, slightly unreal landscapes, somewhere beautiful enough that people wish it existed, then animate them into slow, gentle clips with an image-to-video tool. Pair each with a soft trending instrumental and a short, wistful caption, and keep the whole feed one consistent mood. It's low effort to make and endlessly repeatable, and the calmer and prettier it is, the better it tends to do.

Stuck on a tool or have an AI video question? Reply, I’ll do my best to help.

Created something cool? Send it over and I might feature you. 📺

📫 Want to get your tool featured or advertise to our readers? Let's chat → [email protected]

Thanks for tuning in!

-Kevin