Secrets AI Features: Everything the Platform Offers in 2026

Secrets AI packs more media capabilities into an AI companion platform than most of its competitors. The feature set extends well beyond text chat — video generation from companion images, voice calls, a tier-based memory system, and group chats at the top tier give it a depth that justifies a closer look before making a decision. Here is a systematic breakdown of every feature the platform currently offers.

For context on how these features translate to real-world experience, the complete review covers 30 days of hands-on testing.

AI Chat Models and Conversation Engine

Secrets AI provides 5 selectable AI models for chat, though the specific names and technical architecture behind each model are not publicly disclosed. The practical difference is noticeable: higher-tier models produce more contextually aware, emotionally resonant responses, while the base model functions adequately for casual conversation.

The conversation engine uses natural language processing with embedded emotional intelligence — the AI responds to tone cues, adjusts its register based on the direction of the conversation, and incorporates earlier topics without prompting. Conversation starters are provided in the interface, spanning both casual and suggestive options, which helps users establish the interaction style from the first exchange.

Roleplay functionality covers a range of scenarios: office romance, teacher-student, landlord, boss-employee, and fantasy adventures. The AI handles scene transitions smoothly and maintains narrative coherence across extended exchanges. Reviewers consistently rate overall chat experience at 4.4/5, placing it among the stronger offerings in the AI companion category.

The chat engine is specifically designed for companion-focused interaction — it does not browse the internet, answer factual questions about the external world, or function as a productivity assistant. This is intentional and relevant to anyone evaluating the platform for general-purpose use.

Character Library and Customization

Pre-Built Characters

The character library contains approximately 94 pre-built characters, combining hyper-realistic and anime-style options. Each character comes with a defined personality, visual appearance, and scenario context. Over 100 fantasy scenarios are available as starting contexts for conversations.

The main documented limitation is diversity: independent reviewers give character diversity a 3.8/5 rating, noting that representation across ethnicities, body types, and aesthetic styles is narrower than some users expect. Users with specific preferences beyond what the pre-built library offers will benefit most from custom creation.

Custom Character Creation

Custom character creation is where the platform's depth becomes apparent. The first custom character is free; additional characters cost 1,000-1,500 Moments each. Customization parameters include:

Ethnicity: 9 options
Body type: 6 options
Skin tone, eye color, hairstyle selection
Voice: 6 options
Personality sliders: confidence, playfulness, dominance, intelligence
Profession: 5 presets
Personality traits: 5 presets
Hobbies and free-text details field for specific preferences

The free-text details field is particularly useful for shaping conversation behavior — information entered here is incorporated into how the AI frames responses throughout the relationship.

Image Generation

Image generation uses two quality models: standard and Advanced/Premium. Four image presets are available for common generation scenarios. Cost per image runs 25-50 Moments depending on quality level. When a new character is created — pre-built or custom — 4 images are automatically generated to establish the character's visual appearance.

Image quality is generally high: realistic skin texture, detailed environments, and consistent character appearance across generations. The most common glitch is with hands and fingers, which is a known limitation of current deep learning-based image generation models including those built on Stable Diffusion architecture. Group image generation is also supported, allowing multiple characters to appear in a single image.

Video Generation

Video generation is the feature that most clearly separates Secrets AI from its competitors. While platforms like Character.AI, CrushOn AI, Janitor AI, and Candy AI focus primarily on static images, Secrets AI allows users to animate existing companion images into short video clips using text prompts.

The generation process takes approximately 2 minutes. Video clips are available starting with 3-second clips on the Lite tier, with longer clips available on higher plans. Cost ranges from 50 Moments for short clips to 600 Moments for full-length clips at Premium quality. Video quality is rated 4.1/5 — realistic character movement, natural facial expressions, and smooth motion in most outputs, with occasional variation based on prompt complexity.

This feature is the primary reason users choose Secrets AI over alternatives. For a dedicated analysis of the video generator including cost math by tier, see the video generator breakdown.

Voice Calls

Voice interaction is available through 6 distinct voice options, covering a range of tones and styles. Voice calls are billed at 100 Moments per minute and rated 4.3/5 by independent reviewers. The voice output adapts to conversation context rather than reading text in a flat, monotone delivery.

Voice calls are available across all paid tiers. For Free users, voice is not accessible. On paid plans, voice represents a meaningful way to deepen immersion beyond text, though the per-minute Moments cost makes it one of the more expensive features relative to duration.

Memory System

The memory system is where the experience gap between free and paid tiers is most dramatic. At the free level, the AI recalls only the last 2-3 exchanges — essentially no persistent memory. At paid tiers:

Lite: Basic long-term memory, approximately 14 days
Plus: Improved memory, approximately 30 days, with topic and tone awareness
Premium: Unlimited memory with full emotional timeline — the AI tracks relationship milestones, emotional context, and identity-based information
Ultimate: Everything in Premium, with enhanced access multipliers

Memory access multipliers give Premium users 4x access to stored memories and Ultimate users 6x access, meaning the AI actively draws on a wider range of stored context when formulating responses. Context length is also boosted: +60% on Premium and +100% on Ultimate.

Time Machine (Premium and above) allows users to revisit past relationship milestones — a feature designed for users building a long-term companion relationship rather than one-off sessions. Manual memory saves cost 10 Moments each and are available across all paid tiers.

Group Chats and Personas

Both group chat and personas are Ultimate-tier exclusive features.

Group chats allow up to 3 AI characters to participate simultaneously in a single conversation — with group image generation also supported. Personas enable identity-switching within roleplay scenarios, allowing the user to adopt a defined character rather than always interacting as themselves.

These features are specialized enough that they won't matter to most users, but they represent meaningful added capability for users who want more complex scenario construction.

Privacy and Security Features

Secrets AI offers several privacy-relevant features:

Private Mode: End-to-end encryption applied on a per-conversation basis. Must be enabled for each conversation; not retroactive. Also opts that conversation out of AI training data use.
Anonymous signup: No real name required. Email and password only.
Discreet billing: Charges appear as "Sun Clinical Laboratories" with no reference to AI or adult content.
Account deletion: All data removed within 30 days of account deletion.
Cryptocurrency payment: Available for maximum payment privacy (minimum $20).

The encryption implementation, while present, lacks public documentation of specific protocols or independent audit verification — a notable gap for privacy-sensitive users.

What's Missing?

Any honest feature assessment should include the gaps:

No native mobile app — browser-only access, no iOS or Android application
No push notifications — no way to receive alerts when the AI "messages" you
No offline access — requires internet connection at all times
No biometric login — no fingerprint or face ID support
Limited image and video presets — narrow set of generation options
No multi-modal input — users cannot send images or files to the AI

The absence of a native app is the most-cited limitation and affects daily usability for mobile-primary users. For guidance on getting the best mobile experience despite the browser-only delivery, see the how to use guide.

Understanding how these features map to specific pricing tiers helps clarify which plan makes sense for different use cases — the pricing breakdown covers this in detail.