Secrets AI Video Generator: How It Works, Quality, and Cost
Video generation from AI companion images is Secrets AI's most distinctive technical capability. Character.AI does not offer it. CrushOn AI does not offer it. Janitor AI does not offer it. Candy AI has a limited implementation. Secrets AI's video generator — rated 4.1/5 by independent reviewers — operates at a quality level that puts it ahead of most alternatives in this space. Here is exactly how it works, what it costs, and who should use it.
What the Secrets AI Video Generator Does
The video generator takes an existing AI companion image and animates it using a text prompt describing the desired movement or action. The result is a short video clip of the companion character that reflects both the original image's visual identity and the motion specified in the prompt.
This feature places Secrets AI in a distinct category from most AI companion platforms. Artificial intelligence systems capable of producing coherent, character-consistent motion video from static images require deep learning infrastructure that most companion platforms have not built. The fact that Secrets AI has it — and that the quality holds up to reviewer scrutiny — is a genuine differentiator.
The feature is available on Lite plan and above. Free-tier accounts cannot access video generation. For the full context of what differentiates paid tiers from free, see the free vs premium comparison.
How to Generate a Video — Step by Step
The video creation workflow is straightforward:
- Select a source image: Choose from existing companion images already generated in your account, or generate a new image to use as the video base
- Write a motion prompt: Describe the movement, action, or expression you want — specific prompts produce better results than vague ones (e.g., "slowly turning head and smiling" rather than "moving")
- Submit and wait: The AI processes the request. Standard generation time is approximately two minutes
- Review the output: The completed video clip appears in your session. Download or save as needed
The system uses the character's established visual identity from the source image — hair, face, clothing, body proportions — as the fixed reference, with the prompt controlling motion. Context-aware outputs that reflect the established conversation scenario are possible when the prompt references the current roleplay setting.
Quality Assessment
Independent reviewers rate Secrets AI video quality at 4.1/5. The specific findings from review assessments:
- Character movement: Clips "look good and move smoothly most of the time" — facial expressions, head movement, and upper-body motion show natural-looking results in the majority of outputs
- Visual fidelity: The character's appearance from the source image is preserved with reasonable accuracy across the motion sequence
- Consistency: Natural movements and expressions appear in most outputs; occasional quality variations occur with complex multi-subject prompts
- Generation model impact: The Premium generation model produces noticeably better video quality compared to the standard model. If your tier includes the Premium model option, use it.
The platform is honest that video quality is not perfect — "look good and move smoothly most of the time" implies outliers. Outputs are less reliable when prompts involve complex physical interactions or unusual motion patterns. Starting with simple prompts (head turn, smile, hand movement) before attempting more complex actions is the practical approach to managing this variance.
For direct comparison: Candy AI's video implementation exists but is more limited in scope. No other mainstream AI companion platform reviewed produces video at a comparable feature level.
Moments Cost Breakdown
Video is the most Moments-intensive feature on the platform. The cost structure:
| Video Type | Moments Cost | Notes |
|---|---|---|
| Short clip (3 seconds) | ~50 Moments | Lite tier and above |
| Standard/full clip | ~600 Moments | Higher quality, longer duration |
| Image (for reference) | 25-50 Moments | Source image cost if generating new |
For perspective on the 600-Moment cost for a full video: 600 Moments is equivalent to:
- 12-24 generated images
- 6 minutes of voice calls
- 300-600 text messages
This is a substantial per-use cost. Users who plan to generate video regularly should factor this into plan selection.
Monthly Video Budget by Tier
| Tier | Moments/Month | Short Clips (50 Mo.) | Full Clips (600 Mo.) |
|---|---|---|---|
| Lite ($5.99) | 1,000 | ~20 short clips | 1-2 full clips |
| Plus ($9.99) | 3,000 | ~60 short clips | ~5 full clips |
| Premium ($19.99) | 8,000 | ~160 short clips | ~13 full clips |
| Ultimate ($39.99) | 15,000 | ~300 short clips | ~25 full clips |
The math is clear: Lite is a minimal video budget suited for testing the feature. Plus allows moderate video use if you prioritize it over other Moments consumption. Premium supports regular video creation as part of a mixed media diet. Ultimate is the appropriate tier for users who want video as a primary output.
Top-up Moments bundles are available if you exhaust your monthly allocation — starting at 1,980 Moments for $5.99 up to 118,800 for $249.99 with discounts up to 30.4%. For the full pricing detail, see the Moments cost breakdown.
Video vs Images vs Voice — Where Video Fits
| Feature | Moments Cost | Output | Time Required |
|---|---|---|---|
| Text message | 1-2 | Text response | Seconds |
| Image | 25-50 | Static image | ~30 seconds |
| Short video (3s) | ~50 | Brief motion clip | ~2 minutes |
| Full video | ~600 | Longer motion clip | ~2 minutes |
| Voice call | 100/minute | Real-time audio | Real-time |
For the same 600 Moments, users can choose between: one full-length video, 12-24 images, or 6 minutes of voice. The right trade-off depends on what kind of content you value most. Image generation is the most Moments-efficient visual output. Video offers motion and dynamic character expression that images cannot. Voice provides real-time interactive audio.
Tips for Better Video Results
Practical techniques for improving video output quality and Moments efficiency:
Source image quality matters most. A higher-quality, well-lit source image produces a better animated output. Before spending 600 Moments on a full video, verify that the source image is one of your better generations.
Write specific motion prompts. Vague prompts ("move" or "dance") produce unpredictable outputs. Specific prompts — "slowly turn head to the left, smile" — give the AI clearer parameters and produce more reliable results.
Test with short clips first. At 50 Moments per short clip versus 600 for a full video, testing your prompt on a three-second clip before committing to the full-length version is a rational Moments-conservation strategy. If the short clip quality is good, proceed to the full version.
Use the Premium generation model. The Advanced model on Premium and Ultimate tiers produces meaningfully better video quality. If your subscription includes this option, always select it for video generation.
Save Moments by converting your best images. Rather than generating new images specifically for video, cycle through your existing character images and identify the best ones to animate. This saves the 25-50 Moments per image while still getting good video source material.
Who Should Use the Video Generator
Worth using if:
- Visual content alongside chat is important to your experience
- You want dynamic character media to save or share
- You are on Plus, Premium, or Ultimate with Moments to allocate toward video
- You want a feature that most competing platforms simply do not offer
Not worth prioritizing if:
- Your Moments budget is tight and text chat is your primary use case
- You are on the free tier (video generation is unavailable)
- Image quality is more important to you than motion — Candy AI's still images are sharper
Best tier for video:
- Moderate use: Premium ($19.99/month) — 13 full clips/month while maintaining image and voice budget
- Heavy use: Ultimate ($39.99/month) — 25+ full clips/month
For users who are on the fence about upgrading to access video, the full review covers all features in context of the overall platform value. For a comparison of video access across subscription levels, see the features page.
Try the Secrets AI Video Generator →
Competitors with Video Generation
The AI companion market is sparse on video generation. A survey of the primary competitors:
- Character.AI: No video generation
- CrushOn AI: No video generation
- Janitor AI: No video generation
- Candy AI: Limited video capability, more constrained than Secrets AI
- Replika: No video generation
- SweetDream AI: Video generation available (comparable feature)
- Xotic AI: 4K 15-second clips (higher resolution specification, limited availability)
The competitive landscape makes video generation one of Secrets AI's two strongest differentiators (the other being the tiered memory system). For users whose decision between Secrets AI and an alternative hinges specifically on video capability, the choice is clear — most alternatives do not offer it.
FAQ
Short clips (available at Lite tier and above) run approximately three seconds. Full-length clips are longer — exact duration varies by prompt and generation settings. Lite-tier users access three-second clips; Plus and above unlock longer clip generation. The generation time for any clip is approximately two minutes regardless of clip length.
No. Video generation requires a Lite, Plus, Premium, or Ultimate subscription. Free-tier accounts cannot access the video generator. The free tier includes 200 Moments as a one-time starting allocation, but those Moments cannot be spent on video. Upgrading to Lite ($5.99/month) is the minimum investment required to access the feature.
The monthly video budget depends on your tier and how you split Moments between features. On Plus (3,000 Moments): approximately 5 full-length videos or 60 short clips per month if you allocate all Moments to video. On Premium (8,000 Moments): approximately 13 full-length videos or 160 short clips. On Ultimate (15,000 Moments): approximately 25 full-length videos. Mixed Moments use across images, video, and voice reduces these numbers accordingly.
Independent reviewers rate Secrets AI video quality at 4.1/5. The clips show realistic character movement and facial expressions in most outputs. Character appearance from the source image is preserved with reasonable accuracy. Quality variations occur with complex prompts or multi-subject scenarios. Using the Premium generation model and high-quality source images produces the best results.