Launch your voice AI on X
Voice is the hardest AI to launch with a post. The product is a sound, and a lot of people scroll with the sound off. So a voice launch on X has to make people turn the sound on, and then make them want to try it.
We've launched Cartesia's voice models and Natura Umana's voice-first devices this way. Cartesia's founder post reached 7.1M views, about 340 times the founder's following.
Plan your voice launchVoice AI we've launched on X
Models and devices, heard far past their own audience
The launch came from the founder's own account, with creators behind it on launch day.
Read the case study →The same play, behind Cartesia's own launch post.
Read the case study →Launch day built to feel everywhere at once, across X and LinkedIn.
Read the case study →Why voice launches are different
The demo has to work with the sound off and sound great with it on. Put captions on screen so the clip makes sense muted, and give people a reason to unmute in the first seconds.
Comparison does the explaining. People can't judge a voice in isolation. The same line read two ways, or your model next to what they use today, makes the difference obvious.
Developers save it to build with. Cartesia's founder post earned 1,694 bookmarks. For a voice model, a save usually means someone plans to try the API.
Hardware and models launch differently. A voice model can be tried on launch day. A voice device often can't ship on day one, so the launch needs a different job.
Two kinds of voice launch
A model people can try today, or a device they'll wait for
Get them to try it
- Lead with
- A captioned clip and a comparison
- The one action
- Playground or API link, today
- Who posts
- The founder, then builders who use voice
- Example
- Cartesia, 7.13M views on the founder post
Make it an event
- Lead with
- Someone using it in their day
- The one action
- Waitlist or pre-order
- Who posts
- The founder, then many voices at once
- Example
- Natura Umana, 5.44M views on the launch post
Two kinds of voice launch
Voice models and APIs (like Cartesia). Launch from the founder's own account, with a demo people can hear and a link they can try right away. Creators who build with voice explain what's new and post their own tests. Cartesia used both accounts: the founder post for Sonic-3.5 and Ink-2 reached 7.13M views, and the company post for Sonic-3.6 reached 2.38M, about 150 times its following.
Voice devices (like Natura Umana). Make launch day feel like an event, with many voices posting at the same moment. If people can't buy it today, make the waitlist or pre-order the one clear action. Natura's AI voice launch post reached 5.44M views, about 650 times the founder's following, and its CEO said the waitlist felt it the same day.
How we launch voice AI on X
- Record one clip that works muted. Captions on screen, the best line in the first seconds, and a reason to turn the sound on.
- Show a comparison. The same line in different voices or emotions, or your model next to what people use today.
- Give people something to do. An API key, a playground link, or for hardware, a waitlist or pre-order.
- Launch from the founder. The founder post carries the story. The company account quotes it.
- Line up builders and early adopters. Creators who build with voice, or who try new devices, posting their own tests in the first hours.
- Keep the story going. A follow-up wave on X and a LinkedIn post for business buyers. Natura's CEO's LinkedIn follow-up earned 3,793 reactions.
Demo ideas by voice product
| Voice product | Show this in the demo | Why it works |
|---|---|---|
| Text-to-speech model | One line read in three styles or emotions | People hear the range in seconds |
| Speech-to-text | Fast or accented speech transcribed live | Accuracy is easy to judge on screen |
| Voice agent | A real call from start to finish | Speed and turn-taking are the product |
| Voice device | Someone using it in their day, captioned | People picture it in their own life |
| Voice cloning | A cloned voice, used with the speaker's permission | Shows quality, and shows you take consent seriously |
What we run for voice launches
- Founder launch amplification, with creators behind the founder's own post on launch day.
- Creator tests and demo clips, each creator trying the voice on their own lines.
- Quote posts with each creator's take. Natura's AI voice launch post earned 242 quotes.
- Coordinated launch waves for devices, when launch day needs to feel everywhere at once.
- LinkedIn founder pushes for voice products sold to teams.
You approve every post, creators write in their own words, and paid posts carry X's paid partnership label. Every campaign on X has an agreed impressions floor. If the wave falls short, we add creators at no extra cost until it clears.
A brief template for voice launches
Launch: [PRODUCT NAME] by [COMPANY]
Goes live: [DATE], [TIME AND TIME ZONE]
Try it: [PLAYGROUND, API OR WAITLIST LINK]
What it does: [ONE SENTENCE]
Who it's for: [THE READER YOU BOTH SHARE]
Record one short clip:
- Captions on screen, so it works with the sound off
- Your best line in the first few seconds
- Try it on your own words, not ours
Ideas: [A LINE IN 3 STYLES] / [A COMPARISON] / [A REAL CALL OR TASK]
Send your draft before [TIME] for approval.
Paid posts carry the platform's paid partnership label.Mistakes that sink voice launches
- A demo that needs sound to make sense. Muted scrollers move on. Caption everything.
- Only perfect, hand-picked samples. People trust the voice more when creators test it on their own lines.
- No comparison. Without a reference point, people can't tell what's new.
- Nothing to try. For a model, a playground or API link. For a device, one clear waitlist or pre-order link.
- Cloned voices without permission. Only use a person's voice with their consent, and say so.
Frequently asked questions
How do you market a voice AI product on X?
Lead with a short clip that works with the sound off, show a comparison so people hear what's new, and give them something to try. Launch from the founder's account and line up creators who build with voice to post their own tests in the first hours.
Should a voice AI demo rely on sound?
It should sound great with sound on, but make sense with it off. Put captions on screen and put your best line in the first few seconds.
Should a voice model launch come from the founder or the company?
Both can work. Cartesia's founder post for Sonic-3.5 and Ink-2 reached 7.13M views, and its company post for Sonic-3.6 reached 2.38M.
How do you launch voice AI hardware on X?
Make launch day feel like an event, with many voices posting at the same moment, and give people one clear action like a waitlist or pre-order. Natura's AI voice launch post reached 5.44M views.
What results have your voice AI launches had?
Cartesia's two launch posts passed 9.5M combined views on X. Natura Umana's AI voice launch post reached 5.44M views, about 650 times the founder's following.
Do you run LinkedIn for B2B voice products?
Yes. For voice products sold to teams, we add LinkedIn founder pushes alongside the launch on X. Natura's CEO's LinkedIn follow-up earned 3,793 reactions.