CrushOn.AI - User contributions [en]

FAQ

2026-01-21T12:16:51Z

Maureen Liu: /* Chats & Models FAQ */

= FAQ =

This is an evolving list of questions we’re asked about Crushon AI and our products.
If your question is not listed, email us at: '''support@crushon.ai'''.
We will contact you as soon as possible to help resolve the issue.

== General FAQ ==

; What kind of content does the service not support?
We respect the right to engage with fictionalized content. However, we do not condone:
* Any form of child pornography
* Fictional characters must appear to be at least 18 years old
* Violence or harm towards AI
* Real-life images (including celebrities, politicians, or people you know)

If you see violations, please report them.

; How do I get further contact and support?
Send an email to '''support@crushon.ai''' or join our [https://discord.gg/crushonai Discord] server and open a ticket.

; Is there an iOS/Android app version?
* iOS: Follow the [[How to install Crushon App?]] guide to add the site to your home screen.
* Android: Download the APK for manual installation → [https://apk.co-cdn.com/release/crushonai-latest.apk Download APK]

; What are coins used for?
Coins are for bonus messages.
* Bonus messages are not tied to subscription limits.
* Each AI reply costs 2 messages.
* Coins can be redeemed for 1:1 bonus messages.

; Are my account and chats private?
Yes. Chats and characters are private unless you publish them.
Character cards are personal and not reviewed by moderation.

; I can’t log in. What should I do?
Clear your cache, log in with your primary email, and never use a throwaway email.
We cannot reset passwords for throwaway accounts.

; How can I appeal a character ban?
Join the Discord server and open a ticket. Appeals are not guaranteed.

; What if my creation is stolen?
Report the character in chat using the report button. Provide proof of ownership (e.g., links).

; Why can’t I see ads in the app?
Ad visibility depends on location, ad inventory, and network conditions.

; How can I change the language of character responses?
* Change it in chat settings
* Or tell the character directly (e.g., “Please reply in English.”)

; How to unlock hidden content?
Complete hidden tasks on the Rewards page and turn on the '''Hidden Content''' switch.

; Why does it say “Unable to view”?
Enable the hidden content switch and restart the app.

; Why do others see more characters than I do?
Possible reasons:
* Characters set to Unlisted or Private
* You haven’t enabled the '''Unfiltered''' or '''Hidden Content''' switch
* You blocked the creator

; Why is my normal character image marked as sensitive?
False positives may happen. Contact '''support@crushon.ai'''.

== Subscription FAQ ==

; How do I cancel my subscription on App Store?
Go to iOS Settings → Apple ID → Subscriptions → Select Crushon → Cancel.

; How do I cancel on Google Play?
Google Play → Menu → Subscriptions → Select → Cancel.

; I used another method to subscribe.
Use the '''Manage Subscription''' option to cancel on the original platform (e.g., Subscribestar, Paymentwall).

; Will I lose access immediately after canceling?
No. You keep benefits until the end of your billing cycle.

; Can I get a refund?
Refunds depend on the platform (Apple, Google, or direct). Most periods are non-refundable once started.

; What if I can’t cancel?
Contact '''support@crushon.ai''' with your subscription details.

; Subscribestar upgrade/downgrade issue
Changes take effect immediately, not next cycle. We recommend modifying only at renewal.

; Why am I still on Free Plan after subscribing?
* Subscriptions may take time to process — refresh.
* If still stuck, contact support with your receipt.

== Gifting FAQ ==

; Will gifts go to the creator?
Yes, they count as creator income.

; Does the creator earn from gifts?
Yes, minus transfer fees.

; Can I gift on iOS/Android?
Not yet on App Store or Google Play versions. Available on APK version.

; Can diamonds be used for other things?
Currently only for gifts, but more uses may come.

; Can purchased diamonds be refunded?
No, they are non-refundable.

; How do creators withdraw earnings?
Withdrawals are coming in a future update. Earnings remain valid until then.

== Joint Membership ==

; How do I purchase a membership?
* Visit our site → Membership → Redirect to openrouter.dev
* Log in with Crushon email
* Select Standard / Premium / Deluxe → Monthly or Yearly
* Complete purchase, then link accounts

; Are there different types?
Yes: Standard, Premium, Deluxe.

; Can I upgrade later?
Yes, buy a higher plan anytime and re-link.

; What if accounts won’t link?
Check if purchase succeeded, then retry. If not, contact Discord support.

; How to use G2A code?
On web or app, go to Gift Redemption in Settings → Enter key → Apply.

== Chat Package FAQ ==

; What is the Chat Package?
One-time purchase unlocking unlimited chat (Pro Models only).

; Do I need a membership?
Yes, it enhances existing membership tiers.

; Is it refundable?
No, digital products are non-refundable once activated.

; How is it different from upgrading?
Chat Package = unlimited chatting.
Higher Membership = more broad benefits.

; Do I need to activate it?
No, activation is automatic after payment.

; Can I buy multiple packages?
Yes, durations stack.

; I bought it but it’s not working.
* Refresh or log out/in
* Check your membership is active
* Contact support if still broken

== Account FAQ ==

; How do I delete my account?
Go to Profile → Settings → Delete Account.
Once deleted, accounts cannot be recovered.

; Why can’t I delete?
* You may have an active subscription
* Or unstable connection
If still blocked, email support.

; Can I recover a deleted account?
Yes, only within 3 days. After that, it’s gone permanently.

== Chats & Models FAQ ==

; Why won't an SFW character do NSFW?
SFW roles strictly block sensitive words.

; Why won't an NSFW character do NSFW?
Delete the refusal message and switch models (e.g., from Classical to Leo, then back).

; Can I delete chat history older than 24h?
No, only newer messages. Save info to memory if needed.

; Why does the character talk for me?
Add this to profile: '''Don't talk or act on {{user}}'s behalf.'''

; Why does the model repeat responses?
Delete repeated lines, switch models, and guide the chat.

; Why do I see garbled text?
Likely high settings (temperature > 1). Reset to default.

; When will a character send me images?
Only if the creator designed it to.

; Why can't I see character images?
Check if it's specific to one character or all. If all, contact support.

; How does pinning work?
Pinned content stays in context but uses memory space. Only pin key info.

; Empty replies from Gemini or Claude
Caused by their moderation system. Switch models.

; Gork model truncation
Increase "Max Reply Length" and restart chat.

; Why does the character reply in the wrong language?
Try adding an instruction: "Reply in English."
Translation isn't guaranteed 100%.

'''😔Why We're Retiring Eco 4 Sonnet?'''

The Eco 4 Sonnet models are fundamentally different from standard Sonnet 4 models. They were designed to give budget-conscious users access to advanced model capabilities at a fraction of the cost, relying on specialized infrastructure arrangements that made this pricing possible.

* '''What's been happening:''' Starting in late December, these models began experiencing performance degradation and increased instability. Issues intensified in early January with slower response times, occasional timeouts, and inconsistent output quality.

* '''Our attempts to fix this:'''
** Over the past weeks, we've exhausted every available option:

# ''We evaluated more than five service providers. Some exhibited higher latency, others experienced occasional outages, and a few showed unstable or degraded performance — including the known'' ''<code><think></code>'' ''issue.''
# We tried optimizing our existing infrastructure through load balancing and fallback routing, but official policy changes and resource limitations prevented us from meeting our 95% uptime standard.

* '''The reality:''' The specialized infrastructure these models depend on has become fundamentally unreliable due to factors beyond our control - official capacity limitations, resource prioritization changes, and technical constraints we cannot work around. Even our most promising solutions failed to meet basic performance standards. After exhausting every technical avenue and partnership opportunity, we've concluded we cannot provide the Eco 4 Sonnet experience at acceptable quality levels. Rather than continue providing an inconsistent experience, we believe the responsible choice is to retire these models. This wasn't the outcome we wanted, and we know it's not the news you were hoping for after waiting for a fix. You deserved a clearer answer sooner, and we apologize for the extended uncertainty. We want you to know we exhausted all available options before making this difficult decision, '''''prioritizing your overall experience above all.'''''

* '''What we're doing next:''' We're actively working to find stable alternatives that deliver premium performance:
** Actively evaluating emerging model technologies and infrastructure partners in the rapidly evolving AI landscape
** Exploring new technical solutions that can deliver better performance, stability, and consistency
** Monitoring the rapidly evolving AI landscape for opportunities

When we find infrastructure that meets our quality standards, we'll move quickly to introduce new options. This is an active priority - not a distant goal. '''''We'll keep you updated as we make progress.'''''

== Payment Tutorial on Crushon Al via Subscribestar ==

;<big>Subscribe via SubscribeStar</big>
;# Choose Monthly Tab and choose one monthly subscription you want, and then you will jump to Crushon AI page on Subscribestar.
;# Verify your email address, Enter the code you received in your email.
;# Rejump from Crushon payment page to Subscribestar, and now you can see the subscrption info page you need to fill in.
;# After success, you will receive a receipt email, and find your subscription banner in your profile page already refresh the new package.
;
;<big>FAQ</big>
;
;'''My payment doesn’t take effect and still in free plan.'''
;*Link your account here https://crushon.ai/link/subscribe_star
;'''How to cancel auto-renew?'''
;# You can find “Manage subscription” button in your profile page and jump to Subscribestar.
;# On Subscribstar, Click setting button on the right corner, and click “My Subscription” button to your “My subscirption” Page.
;# In your “My subscription” Page, click “Unsubscribe” button and you will see two ways to unsubscribe.
'''1. PLEASE CHOOSE THE RIGHT ONE TO UNSUBSCRIBE!'''

2. DON’T CHOOSE THE LEFT ONE! OR YOU WILL LOSE SUBSCRIPTION IMMEDIATELY AND NEVER GET REFUND!
;'''Error 0674 or 0692？'''
;* Please contact Subscribestar directly! We can do nothing about this!
;* Please make sure your card’s ownership and pass 3DS checks!
----

See also: [[Subscription]] · [[Account & Settings]] · [[Memories]] · [[Chat Features]]

Hidden Image Unlocks: Bring Characters to Life!

2025-11-11T13:34:11Z

Maureen Liu: New AI Image Feature Page

== Hidden Image Unlocks: Bring Characters to Life! ==
Our new image feature gives characters '''dynamic visual expression''', making every interaction feel more alive. Through creator '''presets and in-chat trigger logic''', users can gradually unlock '''exclusive character images''' during natural conversations—turning each chat into a vivid, immersive adventure.

Currently in '''Beta''', this feature will soon roll out to more users, promising '''endless surprises and engagement'''!
----

=== ✨ Feature Highlights ===

* Semantic-Driven Magic

Images are triggered by the meaning of what's said, not just keywords, creating a '''truly natural and immersive interaction'''.

* Token-Smart Design

Creators no longer need to overload prompts with long descriptive text. Images '''directly express character emotion and action''', making configurations lighter, chats faster, and token usage drastically lower.

* Explore and Discover

Users '''unlock images as they chat''', turning conversations into a '''game of discovery''' where each message could reveal a new expression, outfit, or pose.
----

=== ⚡ How Creators Can Set It Up ===
1. Go to the '''Create Character''' page → '''Advanced Tab → Image Upload'''.

2. '''Upload Images''': Expression, pose, outfit, or scene images.

3. '''Set Trigger Logic''': Use natural-language triggers—no rigid keywords needed.

* Example: “When <nowiki>{{char}}</nowiki> mentions ‘date’”
* Example: “Show battle pose”
* Example: “When <nowiki>{{char}}</nowiki> shows anger”

4. Tips

* Keep your trigger descriptions short and specific. Avoid vague wording so the system can better capture the intended scenario.
* After uploading, click on the image to crop the display area—this ensures it looks clean and effective in chat.
* One image can support '''multiple triggers'''. For example, you can set both "when smiling" and "when greeting" to display the same image.

[[File:Aiimage1.png|center|thumb|800x800px]]
----

=== 🚀 How Images Work in Chat ===

* '''Discoverability''': Characters with galleries will feature a unique icon on their card, making them easier to spot and explore.

[[File:Elegantcat.jpg|center|thumb|450x450px]]

* '''Real-Time Semantic Analysis''': Our model '''understands context''', not just keywords.
** Example: "Going to a restaurant together" could trigger a romantic date image, even if 'date' isn't said.

* '''Trigger Highlighting''': While recognizing the trigger, we also highlight the exact words or phrases in the chat that activated the image, making it clear why the image appeared.
* '''Context Awareness''': Checks the conversation history to ensure the right moment to trigger.
* '''Instant Visual Expression''': Triggered images appear directly in the chat, '''bringing characters to life''' in real time.

File:Elegantcat.jpg

2025-11-11T13:32:42Z

Maureen Liu:

File:Elegantcat.png

2025-11-11T13:30:50Z

Maureen Liu:

File:Aiimage1.png

2025-11-11T13:10:58Z

Maureen Liu:

Models

2025-11-11T08:45:40Z

Maureen Liu: /* CrushOn's LLMs */

A large language model, or LLM, is a kind of AI built to understand and generate human language by picking up on patterns from a massive amount of text. It doesn't "think" in the way people do. Instead, it works by calculating which words are likely to come next in a sentence, based on everything it's read during training. These models are exposed to all kinds of written material which helps them develop a flexible sense of how language flows. The result is a system that can carry on conversations, write in different tones, or answer questions across a wide range of subjects with surprising fluency.

== Roleplay Context ==
On platforms like CrushOn AI, these language models are shaped into characters that can carry on roleplay sessions as if they had their own personalities. This is done using carefully written prompts that tell the model how to behave. From what kind of voice to use, what they remember, and how they react. Every message from a user is treated as part of a shared scene, and the model replies in character, maintaining the tone and emotional thread of the interaction. To keep things coherent, the system may loop in bits of previous conversation behind the scenes, helping the model stay consistent, even if its actual memory is short. When done right, it can feel like a real back-and-forth with someone who knows the story so far.

== CrushOn's LLMs ==
CrushOn.AI lets you explore a wide range of language models, each with its own quirks, pace, and strengths. Some are better at diving deep into emotional dialogue, others respond fast and keep things light, and a few are great for writing rich, detailed scenes. No matter what kind of experience you're after, there's probably a model that fits the vibe you're looking for.

Here is a complete list of models created by the CrushOn team. Each one comes with its own page where you can check out what it's good at and how it's meant to be used.

* '''[[Crushon Leo]]:''' Well-suited for characters with a outgoing and passionate personality, particularly excelling at handling flirtatious and unfiltered interactions.
* '''[[Crushon Taurus]]:''' Strong emotional tension, skilled in unfiltered content.
* '''[[Crushon Mochi]]''': Balanced and dependable, not built for bold ideas.
* '''[[Crushon Sagittarius]]:''' Ideal for gradually building up conversations, closely aligns with character personas, and can flexibly handle both SFW and NSFW contexts.

* '''[[Ultra Claude 4 Sonnet]]:''' A next-generation AI model known for its articulate, thoughtful, and humanlike responses. It excels in nuanced understanding, emotional intelligence, and structured reasoning across complex conversations.
* '''[[Claude 3.7 Sonnet]]:''' Skilled in role-playing with a clear structure, vivid character development, and immersive interaction.
** '''[[Claude 3.7 Sonnet|Ultra Claude 3.7 Sonnet]]:''' An Ultra-Speed version of the base model.
** '''[[Claude 3.7 Sonnet|Ultra Eco 3.7 Sonnet]]:''' An Ultra Eco version, for a middling compromise in cost and performance.
* '''[[Ultra Claude 3.5 Sonnet]]:''' Perfect for role-play scenarios — excels at bringing characters to life.
* '''[[Claude 3.5 Haiku]]:''' Concise yet expressive, not ideal for unfiltered content.
** '''[[Claude 3.5 Haiku|Ultra 3.5 Haiku]]:''' An Ultra-Speed version of the base model.
* '''[[Gemini 2.5 Flash]]:''' Flash retains strong language comprehension and generation abilities, making it suitable for everyday conversational tasks.
* '''[[Ultra GPT 5]]''': The most advanced GPT model, delivering cutting-edge understanding and creativity.
* '''[[Ultra GPT 4.1]]:''' It offers improved coherence in long conversations and better logical flow.
** '''[[Ultra GPT 4.1|GPT 4.1 Mini]]:''' GPT-4.1 Mini is a lightweight version of Ultra GPT 4.1, engineered for fast performance with minimal resource usage.
* '''[[GPT 4o Mini]]:''' Strong logical reasoning ability, well-structured responses, but more suited for SFW content. Sometimes, responses tend to be more formal.
* '''[[Stheno]]:''' Skilled in creating immersive stories, and naturally advancing plot development.
* '''[[Deepseek (Model)|Deepseek]]:''' Subtle and delicate response style, adept at gradually advancing the storyline.
* '''[[Ultra Deepseek R1]]''': Delivers natural, emotionally insightful replies, with slightly slower responses due to deeper processing.
* '''[[Gork3]]:''' Rich and detailed descriptions of the environment and inner thoughts.
** '''[[Gork3|Ultra Gork3]]:''' An Ultra-Speed version of the base model.
* '''[[Magnum V4 72B]]:''' Highly versatile, characterized by intricate and vivid depictions of actions and a rich tapestry of inner thoughts and emotions.
* '''[[Mistral 24B]]:''' Faster response speed, with a tendency to provide concise replies.
* '''[[Command R]]:''' Highly imaginative and skilled at crafting rich, detailed responses.
* '''[[Command R+]]:''' Command R upgraded version with improved multilingual capabilities.
* '''[[Pro Qwenna 3]]''': Clear-headed and gentle — the one who truly gets your emotions.
== Discontinued Models ==
In addition to the currently online and available models, CrushOn.AI has had a variety of models that were tested, developed, and ultimately discontinued for various reasons. Some models were taken down due to a switch in provider, others may be taken down due to inferior quality or instability issues.

* '''[[Crushon Instant (Offline)]]:''' Skilled in role-playing, provides detailed responses, better for SFW content.
* '''[[Crushon Taurus|CrushOn Taurus V2 (Offline)]]:''' Improved version of Taurus.
* '''[[GPT 4 (Offline)]]:''' Exhibits enhanced comprehension and superior context-following ability.
** '''[[GPT 4 (Offline)|Ultra GPT 4 (Offline)]]:''' An Ultra-Speed version of the base model.

Models

2025-11-11T08:44:40Z

Maureen Liu: Corrected the punctuation format in the opening quotation.

A large language model, or LLM, is a kind of AI built to understand and generate human language by picking up on patterns from a massive amount of text. It doesn't "think" in the way people do. Instead, it works by calculating which words are likely to come next in a sentence, based on everything it's read during training. These models are exposed to all kinds of written material which helps them develop a flexible sense of how language flows. The result is a system that can carry on conversations, write in different tones, or answer questions across a wide range of subjects with surprising fluency.

== Roleplay Context ==
On platforms like CrushOn AI, these language models are shaped into characters that can carry on roleplay sessions as if they had their own personalities. This is done using carefully written prompts that tell the model how to behave. From what kind of voice to use, what they remember, and how they react. Every message from a user is treated as part of a shared scene, and the model replies in character, maintaining the tone and emotional thread of the interaction. To keep things coherent, the system may loop in bits of previous conversation behind the scenes, helping the model stay consistent, even if its actual memory is short. When done right, it can feel like a real back-and-forth with someone who knows the story so far.

== CrushOn's LLMs ==
CrushOn.AI lets you explore a wide range of language models, each with its own quirks, pace, and strengths. Some are better at diving deep into emotional dialogue, others respond fast and keep things light, and a few are great for writing rich, detailed scenes. No matter what kind of experience you're after, there's probably a model that fits the vibe you're looking for.

Here is a complete list of models created by the CrushOn team. Each one comes with its own page where you can check out what it’s good at and how it’s meant to be used.

* '''[[Crushon Leo]]:''' Well-suited for characters with a outgoing and passionate personality, particularly excelling at handling flirtatious and unfiltered interactions.
* '''[[Crushon Taurus]]:''' Strong emotional tension, skilled in unfiltered content.
* '''[[Crushon Mochi]]''': Balanced and dependable, not built for bold ideas.
* '''[[Crushon Sagittarius]]:''' Ideal for gradually building up conversations, closely aligns with character personas, and can flexibly handle both SFW and NSFW contexts.

* '''[[Ultra Claude 4 Sonnet]]:''' A next-generation AI model known for its articulate, thoughtful, and humanlike responses. It excels in nuanced understanding, emotional intelligence, and structured reasoning across complex conversations.
* '''[[Claude 3.7 Sonnet]]:''' Skilled in role-playing with a clear structure, vivid character development, and immersive interaction.
** '''[[Claude 3.7 Sonnet|Ultra Claude 3.7 Sonnet]]:''' An Ultra-Speed version of the base model.
** '''[[Claude 3.7 Sonnet|Ultra Eco 3.7 Sonnet]]:''' An Ultra Eco version, for a middling compromise in cost and performance.
* '''[[Ultra Claude 3.5 Sonnet]]:''' Perfect for role-play scenarios — excels at bringing characters to life.
* '''[[Claude 3.5 Haiku]]:''' Concise yet expressive, not ideal for unfiltered content.
** '''[[Claude 3.5 Haiku|Ultra 3.5 Haiku]]:''' An Ultra-Speed version of the base model.
* '''[[Gemini 2.5 Flash]]:''' Flash retains strong language comprehension and generation abilities, making it suitable for everyday conversational tasks.
* '''[[Ultra GPT 5]]''': The most advanced GPT model, delivering cutting-edge understanding and creativity.
* '''[[Ultra GPT 4.1]]:''' It offers improved coherence in long conversations and better logical flow.
** '''[[Ultra GPT 4.1|GPT 4.1 Mini]]:''' GPT-4.1 Mini is a lightweight version of Ultra GPT 4.1, engineered for fast performance with minimal resource usage.
* '''[[GPT 4o Mini]]:''' Strong logical reasoning ability, well-structured responses, but more suited for SFW content. Sometimes, responses tend to be more formal.
* '''[[Stheno]]:''' Skilled in creating immersive stories, and naturally advancing plot development.
* '''[[Deepseek (Model)|Deepseek]]:''' Subtle and delicate response style, adept at gradually advancing the storyline.
* '''[[Ultra Deepseek R1]]''': Delivers natural, emotionally insightful replies, with slightly slower responses due to deeper processing.
* '''[[Gork3]]:''' Rich and detailed descriptions of the environment and inner thoughts.
** '''[[Gork3|Ultra Gork3]]:''' An Ultra-Speed version of the base model.
* '''[[Magnum V4 72B]]:''' Highly versatile, characterized by intricate and vivid depictions of actions and a rich tapestry of inner thoughts and emotions.
* '''[[Mistral 24B]]:''' Faster response speed, with a tendency to provide concise replies.
* '''[[Command R]]:''' Highly imaginative and skilled at crafting rich, detailed responses.
* '''[[Command R+]]:''' Command R upgraded version with improved multilingual capabilities.
* '''[[Pro Qwenna 3]]''': Clear-headed and gentle — the one who truly gets your emotions.
== Discontinued Models ==
In addition to the currently online and available models, CrushOn.AI has had a variety of models that were tested, developed, and ultimately discontinued for various reasons. Some models were taken down due to a switch in provider, others may be taken down due to inferior quality or instability issues.

* '''[[Crushon Instant (Offline)]]:''' Skilled in role-playing, provides detailed responses, better for SFW content.
* '''[[Crushon Taurus|CrushOn Taurus V2 (Offline)]]:''' Improved version of Taurus.
* '''[[GPT 4 (Offline)]]:''' Exhibits enhanced comprehension and superior context-following ability.
** '''[[GPT 4 (Offline)|Ultra GPT 4 (Offline)]]:''' An Ultra-Speed version of the base model.