xAI launches Grok Imagine with audio-enabled video and a spicy mode

While most AI labs are busy adding more safety guardrails to their media generators, Elon Musk is going full throttle in the opposite direction.

xAI has officially rolled out Grok Imagine, its new AI image and video generator, to all SuperGrok and Premium+ subscribers on iOS via version 1.1.33 of the Grok app. The tool can turn text or image prompts into 15-second stylized videos complete with native audio in just seconds.

True to Musk's ethos of building an unfiltered AI, the biggest talking point is its built-in "spicy mode," which allows users to generate explicit NSFW content right inside the app.

Fast clips, waxy faces, and a spicy toggle

Speed is Grok Imagine's main flex. Musk claimed on X that the model can generate a full video in one-quarter to one-half the time rivals take to generate a single still image. The interface is remarkably fluid: it generates images instantly from text prompts, continues auto-generating as you scroll down, and lets you animate those frames into videos with a tap.

Then there is the content boundary. Following last month's debut of an anime companion, CNBC recently reported on a deleted post from xAI staffer Mati Roy mentioning a spicy mode capable of nudity. In hands-on testing, TechCrunch confirmed the model can produce semi-nude imagery, though overly explicit prompts still hit moderation filters that return blurred outputs.

The tool also permits generating public figures like Donald Trump and Taylor Swift, though edge cases like trying to create a pregnant Trump failed, yielding him holding a baby instead. In terms of aesthetic quality, Grok Imagine still lingers in the uncanny valley, often producing cartoonish or waxy skin tones rather than the photorealistic cinematic sheen of Google Veo 3.

Why it matters

For creators and casual users, Grok Imagine is basically a speed demon built for social feeds. If you need quick memes, stylized motion clips, or content without constantly tripping puritanical safety filters, it is genuinely fun and fast. But if you need polished commercial footage, the waxy faces and jittery textures mean it is definitely not ready to replace traditional production tools.

For the broader industry, xAI is playing a dangerous game of asymmetric warfare. While Google, Runway, and OpenAI bend over backward to appease enterprise risk teams and regulators, Musk is leveraging loosened restrictions and raw speed as primary hooks to drive premium X subscriptions.

We think this play is clever for short-term growth, but it comes with massive liability. Combining celebrity likeness generation, native sound, and relaxed NSFW filters in a tool that renders clips in seconds is practically a magnet for deepfake scandals. Musk says the model will get better every day, but keeping it out of legal crosshairs might prove much harder than teaching it to animate.