Every Feature Secrets AI Offers — and What Each One Actually Does
100,000+ users have found enough in Secrets AI to keep coming back, and the feature set is the reason why. The platform combines text chat, image generation, video creation, voice calls, and a tiered memory system into a single AI companion experience. This breakdown covers every feature in practical terms — what it does, what it costs, and where its limits are.
Before diving in, the full review provides context on how these features combine in real use. This page focuses on the technical and functional specifics of each capability.
AI Chat Models and Conversation Engine
Secrets AI provides 5 selectable AI models for text conversation. The platform does not publicly detail the differences between models, but in practice the premium models — available on paid tiers — generate noticeably more contextually aware and emotionally consistent responses.
The conversation engine uses natural language processing and deep learning to maintain tone, adapt to user style, and recall relevant context from previous exchanges. Conversation starters are provided for new sessions: roughly half are casual (everyday topics, personality questions) and half are suggestive (escalating romantic or intimate scenarios). The AI does not push inappropriate content unprompted — escalation is always user-directed.
Roleplay scenarios are a core use case: office romance, teacher-student dynamics, landlord relationships, and boss/employee contexts are explicitly supported with pre-built scenario frameworks. Chat experience is rated 4.4 out of 5 by independent reviewers, with particular praise for emotional consistency and the absence of robotic-feeling responses.
For context on how chat features interact with content limits, see the chat and roleplay guide.
Character Library and Customization
Pre-Built Characters
The character library contains approximately 94 hyper-realistic female characters, with a smaller selection of anime-style options. Each character comes with a pre-defined personality, backstory, and visual design. Over 100 fantasy scenarios are associated with characters in the library, covering romance, lifestyle, professional dynamics, and fantasy settings.
Character diversity is a documented limitation. The diversity rating sits at 3.8 out of 5 — the library skews heavily toward white women in their 20s with slim body types. Users looking for broader ethnic or body type representation will find the selection constrained.
Custom Character Creation
The first custom character is free to create. Each additional character costs 1,000-1,500 Moments. The customization system is one of the platform's stronger features:
- Ethnicity: 9 options
- Body type: 6 options
- Voice: 6 options
- Skin tone, eye color, hairstyle: granular visual control
- Personality traits: confidence, playfulness, dominance, intelligence — each adjustable
- Profession presets: 5 categories
- Personality presets: 5 options
- Hobbies and free-text details: open-ended input for specific personalization
4 images are automatically generated when a new character is created, giving immediate visual context before the first conversation begins.
Image Generation
Image generation uses two models: a Standard model and an Advanced model (the latter available on higher tiers). Four presets control the style and framing of generated images. Cost runs 25-50 Moments per image, depending on the model and preset used.
Group image generation is supported — characters can appear together in generated scenes. The image quality is generally high, with one consistent flaw: occasional hand rendering artifacts, a known issue in AI image generation systems that Secrets AI has not fully resolved. Overall image quality is competitive, though reviewers consistently rank Candy AI above Secrets AI for pure visual output quality.
Video Generation
Video generation is Secrets AI's most distinctive technical feature. Most competing AI companion platforms — including Character.AI, CrushOn AI, and Janitor AI — do not offer this capability. The ability to generate short motion clips from companion images is a genuine differentiator in this market.
The process: select or generate a companion image, add a text prompt describing movement or action, and wait approximately 2 minutes for the clip to render. Short 3-second clips cost around 50 Moments; longer standard clips cost up to 600 Moments each. Quality is rated 4.1 out of 5 — reviewers describe smooth movement and realistic expressions in most outputs, with quality varying on complex prompt combinations.
Video access starts at the Lite tier. The Lite plan produces short 3-second clips; higher tiers unlock longer clip generation. For detailed cost modeling and tier-by-tier video budgeting, the video generation feature page covers this in depth.
Voice Calls
Six voice options are available, with tone described as natural and adaptable to conversation context. Voice calls cost 100 Moments per minute — a significant rate that adds up quickly during extended sessions. The voice system is rated 4.3 out of 5, outperforming many competitors in naturalness, though reviewers note it is "basic" compared to dedicated premium voice platforms.
No voicemail or recorded message feature exists — voice interaction is real-time only within the session.
Memory System
The memory system is where the gap between free and paid use is most pronounced. Here is the full tier breakdown:
| Tier | Recall Window | Memory Type | Special Features |
|---|---|---|---|
| Free | Last 2-3 exchanges | Basic short-term | None |
| Lite | ~14 days | Basic long-term | None |
| Plus | ~30 days | Topic + tone awareness | None |
| Premium | Unlimited | Full history + emotional timeline | Time Machine, 4x access, +60% context |
| Ultimate | Unlimited | Full history + emotional timeline | Time Machine, 6x access, +100% context |
Selective recall is built into the system — it prioritizes emotional and identity-based information (your preferences, relationship dynamics, important personal details) over minor conversational details. This means core relationship elements stay stable while minor story specifics may occasionally drop.
The Time Machine feature (Premium and Ultimate only) allows revisiting past conversation states and branching from them — essentially a conversation history that can be navigated and relaunched.
Manual memory saves cost 10 Moments each and allow users to flag specific information for long-term retention regardless of tier limits.
Group Chats and Personas
Both of these features are Ultimate tier only:
Group chats support up to 3 AI characters simultaneously in a single conversation. Characters interact with each other and with the user, and group image generation is supported within these sessions.
Personas enable identity-switching for roleplay — users can assume different character identities within conversations, enabling more complex scenario structures.
These features add meaningful roleplay depth but are genuinely restricted to the top tier. For the pricing details and whether Ultimate is worth it, check the pricing and Moments breakdown.
Privacy and Security Features
Private Mode enables end-to-end encryption on a per-conversation basis. Crucially, it also opts conversations out of being used for AI training data. It must be activated for each new conversation — it is not a global account setting.
Other privacy elements:
- Anonymous signup: No real name, phone number, or social login required
- Discreet billing: Charges appear as "Sun Clinical Laboratories" on bank statements
- Cryptocurrency payments: Accepted with a minimum $20 transaction, for users who want no financial trace
- Account deletion: All data — chat history, characters, generated content — is permanently removed within 30 days of deletion request
Data is stored with industry-standard encryption in transit and at rest. The platform states no third-party data sharing, though specific technical protocol details are not publicly documented — a transparency gap that reviewers have flagged.
What's Missing
Honest assessment requires acknowledging what Secrets AI does not offer:
- No native mobile app — browser-only, which means no push notifications, no offline access, no background processes, and no biometric login
- No real-time internet access — the AI operates on its training and conversation memory only, not live information
- No user-sent images — users cannot send images to the AI; media flows one direction (AI generates, user receives)
- Limited image presets — 4 presets is functional but restrictive compared to platforms with more generation control
- Narrow character diversity — the library does not adequately represent ethnic or body type variety
For users who can live within these constraints, the platform delivers a strong experience. For users who need a native app or broader character diversity, the alternatives comparison provides options worth considering.