Aug X released Augie, an AI-powered video creation platform incorporating a voice cloning feature to read ad copy without booking a recording studio.
Partnering with ElevenLabs, Aug X lets users record their or someone else’s voice and clone it to use for other short videos. The Augie platform, aimed primarily toward marketers and social media teams, lets people quickly add narration, photos, text, and music to videos without needing to learn audio and video editing.
Jeremy Toeman, founder of Aug X, said the company wanted to add a voice cloning feature after realizing some people do not like speaking into a microphone or recording voice-overs.
“You’d be surprised at the number of people who’ve used our voice cloning feature that came up to us and says thank you because they didn’t like recording narration,” he says.
In a demo to The Verge, Toeman says users can either write or upload a script to Augie and then use a prerecorded voice (people need to record a short snippet of their voice saying anything) or choose one from its library. That voice will then read the script with the tone, serious, enthusiastic, creepy, etc., which can be adjusted depending on the mood of the video.
Users can also choose from a library of photos — Aug X licenses photos from Getty — or use AI-generated images to add to the video.
Toeman envisions marketing teams turning around short videos without the need to book their company spokesperson time in the recording studio.
Augie went on public beta in May. For now, everyone using Augie has access to the voice cloning feature. Toeman says eventually, it might limit video length and quality and the number of cloned voices of non-paying subscribers because of the expense of running voice cloning servers with its partners.
Voice cloning is not new; Samsung even lets people use an AI voice to respond to calls, and now companies like Aug X are integrating the technology into more full-featured services. After a deepfaked Drake song featuring an AI-generated version of the rapper’s voice went viral, it helped reignite concerns about copyright and illegally using someone’s else likeness.
“We’ve pre-thought what could go wrong, so we’re being really careful about who can use the pre-recorded voice to clone,” Toeman says.
He added that recorded voices, as opposed to the ones part of the larger Augie library, will only be available to individual accounts. So another person in the same organization has to rerecord their spokesperson’s voice on their Augie account to clone their speech. Users cannot upload recordings, and the platform will only accept voices recorded with a live mic. So, no putting your ex’s voice on speaker so you can fake a call with a weirdly helpful friend won’t work with Augie.
The company is also working with its voice cloning partner ElevenLabs to identify AI-generated voices and videos.