xAI added per-request content guidelines to the Grok API on October 27, letting developers set their own rules for NSFW content, violence, and profanity in text API requests. The update gives adult AI and companion developers a documented moderation control instead of a gray-area jailbreak — and it lands two weeks after Grok's consumer app cut off sexual content for UK users under pressure from Ofcom.
What changed
Coverage of the update reports that xAI shipped per-request content guidelines on the Grok API, allowing developers to define their own NSFW, violence, and profanity rules. In xAI's own words, developers "can now set content guidelines on text API requests," passing a guidelines parameter that tells the model how much explicit content to allow in a given response.
That is a meaningful shift for the NSFW AI market. Most frontier API providers enforce a single platform-wide content policy that developers cannot override, which is why adult companion builders have historically either routed around guardrails or moved to open-weight models. Per-request guidelines make moderation a per-call engineering decision, closer to how the adult industry already runs its own infrastructure.
The UK contrast
The developer-facing loosening sits awkwardly next to what happened on the consumer side. On October 14, UK regulator Ofcom said SpaceXAI — the company formerly known as xAI — added new restrictions to its standalone Grok app after talks with the regulator. People in the UK can no longer view, generate, or share pornographic content through Grok's image tools, and Grok's companion chatbots, including Ani, stopped engaging in sexually explicit conversations with UK users.
Together, the two moves frame xAI's position: the API is treated as a developer tool with configurable boundaries, while the consumer app is being pulled toward stricter compliance in regulated markets. For companion developers, that split matters. A model that refuses explicit chat in one jurisdiction can still be configured to allow it elsewhere via the API — but only if the developer takes responsibility for the moderation settings on each request.
What it means for NSFW AI builders
- Sanctioned moderation. NSFW AI developers on Grok get a documented path to set content rules per request rather than relying on jailbreaks that platforms can patch overnight.
- Regional divergence. The UK block shows consumer-facing companion features remain the most exposed to regulator pressure, even as API access loosens.
- Compliance homework. Companion apps routing through Grok now need to track both API guidelines and regional restrictions, since the two can diverge sharply.
The October 27 API update is a rare example of a frontier lab handing NSFW moderation decisions to developers. Whether that survives contact with regulators — and with the UK precedent — will tell companion builders how much of Grok's API openness they can actually build on.