Quick answer
- An NSFW filter is a moderation layer that classifies or blocks sexually explicit text or images, in what you send or what the model replies.
- "NSFW filter" is colloquial. Each app writes its own policy, and app stores separately require filtering of objectionable content.
- Official wording differs: Character.AI calls its filters final, CrushOn.AI has three content tiers, Janitor AI allows adult content between consenting adult characters, and Kindroid says it applies no filter to the AI itself.
- "Unfiltered" never means rule-free: every app we checked publishes banned content.
What Is an NSFW Filter?
An NSFW filter is a content-moderation layer that classifies or blocks sexually explicit material in what a user sends or what a language model returns. It is a loose label; model providers talk about categories and thresholds: Google's Gemini API lets developers "adjust these settings across filter categories to restrict or allow certain types of content", with "Sexually explicit" as one category. OpenAI's moderation models return results an app can use for "filtering content, routing a request for review, or intervening with accounts that submit flagged content."
A companion app's filter is rarely one switch. It can be a model refusing a request, a classifier screening text and images, a tag or toggle you choose, or an account-level scan followed by a warning or lock. CrushOn.AI's wiki adds that some third-party models can return empty replies "caused by their moderation system".
NSFW Filter Policies in Named Apps
Character.AI. The help center says pornographic content "will not be supported at any point in the future" and calls its stance on NSFW filters "final". Asking to remove them "will result in a ban". The Creator Guide says filtering "happens at the platform level" and "isn't a Definition dial you can turn". The company is in a court-supervised sale, so recheck the policy (Chapter 11 report).
CrushOn.AI. The official wiki describes three tiers. Filtered, the default, allows no sexual content in chats or images. Unfiltered lets characters engage in sexual roleplay, but images cannot show nudity or sexual acts. Hidden Content gives registered users full NSFW access after logging in, checking in for two days and spending 50 message credits.
Janitor AI. Its Community Guidelines (February 4, 2026) say it "permits adult content between consenting adult characters". Characters with sexual or romantic themes carry the Limitless tag, and Limited means none. Some categories are banned whatever the tag, including pornographic or nude images. The mobile apps start in a Restricted Mode you lift on the website.
- Kindroid: "we are not implementing a filter on the AI itself"; automated account-level scans and warn-first locking instead.
- Nomi: unfiltered text, but Nomis are not permitted to take nude photos.
- SpicyChat: a profile toggle shows or hides NSFW chatbots and blurs explicit images.
App Store Rules Behind Mobile Filters
Stores add their own layer. Apple's guideline 1.2 requires apps with user-generated or AI content to include "a method for filtering objectionable material from being posted to the app", and a web service's incidental NSFW content may appear only if it is hidden by default and enabled through the website. Google Play's user-generated content policy says incidental sexual content "must be filtered by default and restricted from minors". That is why mobile versions are often stricter than websites; see Apple and Google's tighter AI companion rules.
Related Pages
For which apps allow what, read unfiltered AI chat. Adult romance and images are in the NSFW AI girlfriend ranking, free options in free NSFW AI chat; more terms are in the glossary.
FAQ
Can you turn off an NSFW filter?
It depends on the app. Character.AI says no, for every user including adults. CrushOn.AI has Unfiltered and Hidden Content switches, and Janitor AI uses content tags and a mobile Restricted Mode.
Is an NSFW filter the same as an age check?
No. A filter decides what content is allowed; an age check decides who may see it. CrushOn.AI relies on a self-declared 18+ pop-up, while SpicyChat and Janitor AI require verification in certain regions.
Sources
Official pages and store listings we read for this page (last check October 20, 2026).
- ai.google.dev: Gemini adjustable safety filters incl. 'sexually explicit' category…
- platform.openai.com: OpenAI moderation guide (filtering use)
- developer.apple.com: Apple App Review Guidelines full text (Last Updated Oct 16, 2026)
- support.google.com: Google Play User Generated Content policy
- support.character.ai: NSFW stance
- support.character.ai: NSFW filters stance
- aiwiki.crushon.ai: Official wiki: Filtered / Unfiltered / Hidden content tiers
- janitorai.com: Community Guidelines (2026-02-04)
- janitorai.com: Mobile moderation post
- kindroid.ai: Help center: moderation guidelines
- wiki.nomi.ai
- docs.spicychat.ai: FAQ (NSFW/SFW, models)