Captions, an AI-powered video creation and editing app for contented creators that has secured complete $100 cardinal successful task superior to day astatine a valuation of $500 million, is rebranding to Mirage, nan institution announced connected Thursday.
The caller sanction reflects nan company’s broader ambitions to go an AI investigation laboratory focused connected multimodal foundational models specifically designed for short-form video contented for platforms for illustration TikTok, Reels, and Shorts. The institution believes this attack will separate it from accepted AI models and competitors specified arsenic D-ID, Synthesia, and Hour One.
The rebranding will besides unify nan company’s offerings nether 1 umbrella, bringing together nan flagship creator-focused AI video platform, Captions, and nan precocious launched Mirage Studio, which caters to brands and advertisement production.
“The measurement we spot it, nan existent title for AI video hasn’t begun. Our caller identity, Mirage, reflects our expanded imagination and committedness to redefining nan video category, starting pinch short-form video, done frontier AI investigation and models,” CEO Gaurav Misra told TechCrunch.

The income transportation down Mirage Studio, which launched successful June, focuses connected enabling brands to create short advertisements without relying connected quality talent aliases ample budgets. By simply submitting an audio file, nan AI generates video contented from scratch, pinch an AI-generated inheritance and civilization AI avatars. Users tin besides upload selfies to create an avatar utilizing their likeness.
What sets nan level apart, according to nan company, is its expertise to nutrient AI avatars that person natural-looking speech, movements, and facial expressions. Additionally, Mirage says it doesn’t trust connected existing banal footage, sound cloning, aliases lip-syncing.
Mirage Studio is disposable nether nan business plan, which costs $399 per period for 8,000 credits. New users person 50% disconnected nan first month.
Techcrunch event
San Francisco | October 27-29, 2025
While these devices will apt use brands wanting to streamline video accumulation and prevention immoderate money, they besides spark concerns astir nan imaginable effect connected nan imaginative workforce. The increasing usage of AI successful advertisements has prompted backlash, arsenic seen successful a caller Guess ad successful Vogue’s July people version that featured an AI-generated model.
Additionally, arsenic this exertion becomes much advanced, distinguishing betwixt existent and deepfake videos becomes progressively difficult. It’s a difficult pill to swallow for galore people, particularly fixed really quickly misinformation tin dispersed these days.
Mirage precocious addressed its domiciled successful deepfake exertion successful a blog post. The institution acknowledged nan genuine risks of misinformation while besides expressing optimism astir nan affirmative imaginable of AI video. It mentioned that it has put moderation measures successful spot to limit misuse, specified arsenic preventing impersonation and requiring consent for likeness use.
However, nan institution emphasized that “design isn’t a catch-all” and that nan existent solution lies successful fostering a “new benignant of media literacy” wherever group attack video contented pinch nan aforesaid captious oculus arsenic they do news headlines.