Secrets AI Features: Everything the Platform Offers in 2026

Five AI models, video generation from companion images, six-voice call support, and a tiered memory system that scales from basic recall on free to full emotional timelines on Premium — Secrets AI packs more feature density than most AI companion platforms in its price range. This page maps every capability the platform offers, with honest notes on where each feature works well and where it falls short.

For context on whether these features justify the cost, see the full platform review.

AI Chat Models and Conversation Engine

The chat system runs on five selectable AI models. Secrets AI does not publicly document which model architectures are in use or the specific differences between the five options — users discover differences through experience rather than documentation. In practice, the models vary in response nuance and emotional range, with higher-tier models producing more contextually aware and emotionally consistent outputs.

Conversation quality as measured by external reviewers sits at 4.4/5. The AI handles a wide range of tones — casual small talk, emotional support, suggestive exchanges, and full narrative roleplay — without defaulting to robotic filler phrases. It asks follow-up questions, revisits earlier topics, and tracks the user's stated preferences and interests within a session.

The platform provides conversation starters for each character — approximately half casual, half suggestive — so users do not need to initiate from scratch. The AI adapts to the user's communication style within a few exchanges, leaning playful if the user is playful, or more emotionally substantive if the user engages that way.

Supported roleplay scenarios include over 100 pre-built fantasy contexts: office romance, teacher, landlord, boss, fantasy adventure settings, everyday relationship scenarios, and more. Personalized scenario creation is also supported through the character's free-text details field.

Character Library and Customization

Pre-Built Characters

The pre-built library contains approximately 94 hyper-realistic AI characters. A smaller subset of anime-style characters is also available. Each character comes with a defined personality, backstory, appearance, and voice assignment. Scenario variety across the library is broad — the 100+ pre-built scenario templates cover a wide range of relationship dynamics and fantasy contexts.

The main limitation is visual diversity. The library skews toward a narrow range of appearances, and underrepresentation of certain ethnicities and body types is a documented weakness. Reviewers rate character diversity at 3.8/5 — the platform's lowest-scoring feature area.

Custom Character Creation

Custom character creation addresses the library's diversity gaps to some extent. The customization parameters cover:

• Ethnicity: 9 options
• Skin tone: adjustable
• Eye color: adjustable
• Hairstyle: multiple options
• Body type: 6 options
• Voice: 6 options

Personality is defined across four axes: confidence, playfulness, dominance, and intelligence. Five profession presets and five personality trait presets are available, plus free-text hobby and detail fields for specific customization that presets cannot capture.

The first custom character is free. Additional custom characters cost 1,000–1,500 Moments each.

Image Generation

Image generation uses two models: standard (available from Plus and above) and Advanced/Premium (available on Premium and Ultimate). Each new character automatically generates four images during setup. Additional images can be requested at any time from within a conversation at a cost of 25–50 Moments per image. Four image presets are available to guide style and framing.

Quality for realistic characters is consistently strong. Minor artifacts appear occasionally — hands are the most common issue, which is a known limitation of current AI art generation methods based on diffusion model outputs. Group image generation, involving multiple AI characters in a single image, is supported.

Video Generation

Video generation is Secrets AI's most distinctive feature. Users generate short animated video clips from existing companion images by writing a text prompt describing the movement or action they want. The deep learning pipeline processes the request in approximately two minutes before delivering the completed clip.

Clip length varies by tier: Lite users get 3-second clips, while higher tiers unlock longer outputs. Quality is rated 4.1/5 — character movement and facial expressions are realistic in most outputs, with occasional variation depending on prompt complexity. Using higher-quality source images and the Premium generation model improves results meaningfully.

Cost structure: short clips (~3 seconds) cost approximately 50 Moments; full-length clips cost up to 600 Moments. Video is the most Moments-intensive feature on the platform.

For a complete guide to the video generation feature including cost math and competitor comparison, see the video generator page.

Voice Calls

Six voice options are available for real-time voice interaction. The voices are natural-sounding and tonally adaptable — they modulate between playful, warm, and more serious tones based on conversation context rather than staying flat throughout. Voice quality earns a 4.3/5 reviewer rating.

Cost: 100 Moments per minute. At Premium tier (8,000 Moments/month), this allows approximately 80 minutes of voice time — a reasonable allocation if the user mixes voice with other features. Lite users (1,000 Moments/month) have roughly 10 minutes of voice available, which is tight.

Memory System

The memory system is the feature most directly affected by subscription tier. The functional differences between tiers are significant:

The emotional timeline feature on Premium maintains a structured record of the relationship's development, allowing the AI to reference past emotional events coherently rather than just surface-level facts. The Time Machine feature allows users to revisit previous conversation states.

Manual memory saves are available at 10 Moments per entry — useful for ensuring specific important conversation details are retained regardless of tier.

For detailed analysis of how memory affects value across tiers, see the how to use guide.

Group Chats and Personas

Two exclusive Ultimate-tier features expand the social simulation capability:

Group chats: Engage up to three AI characters simultaneously in a single conversation thread. Characters respond to each other and to the user, creating more complex social dynamics than single-companion interactions.

Personas: Identity-switching capability for roleplay scenarios where the user takes on a different name, background, or personality within a conversation. Useful for structured roleplay that requires character differentiation.

Both features are unavailable below Ultimate ($39.99/month). Group image generation — producing a single image containing multiple characters — is also supported alongside group chat.

Privacy and Security Features

• Private Mode: End-to-end encryption applied on a per-conversation basis. Must be enabled manually for each conversation. When active, the conversation is excluded from AI training data.
• Anonymous signup: Email and password only — no real name, phone number, or social media login required.
• Discreet billing: Charges appear on bank statements as "Sun Clinical Laboratories," with no reference to AI services or adult content.
• Cryptocurrency payments: Accepted with a minimum of $20, enabling maximum payment privacy.
• Account deletion: All data removed within 30 days of deletion request.

For a complete safety and privacy assessment, visit the is it safe page.

What's Missing?

No feature set is complete without honest gaps. Secrets AI currently lacks:

• Native mobile app: No iOS or Android application. Browser-only access.
• Push notifications: No alerts when a companion sends a message (because sessions are user-initiated).
• Offline access: Requires internet connection for all features.
• Image input: Users cannot send photos to the AI — media flow is one-directional (platform generates for user).
• Expanded character diversity: The pre-built library's narrow visual range is a recurring complaint.
• Limited presets: Image and video generation preset options are constrained compared to dedicated AI art platforms.

These gaps matter differently depending on user priorities. For users who primarily want text conversation and occasional media generation through a desktop browser, none of them are blocking issues. For mobile-first users, the lack of a native app with push notifications is a genuine limitation.

For pricing and Moments details, visit the pricing plans page.