Secrets AI Video Generator: How It Works, Quality, and Cost
Most AI companion platforms generate images. A small number of them generate video. Secrets AI sits in that second group — and the gap matters. Turning a static companion image into a motion clip represents a genuinely different capability from still image generation, and it is the feature that most clearly distinguishes Secrets AI from Character.AI, CrushOn AI, Janitor AI, and most other platforms in this category.
This page covers exactly how the video generator works, what the quality looks like in honest terms, the full Moments cost math, and who should (and should not) prioritize this feature.
What Is the Secrets AI Video Generator?
The video generator converts a static AI companion image into a short motion clip using a text prompt describing the desired movement or action. The generated video reflects the character's appearance, the scene context, and the movement specified in the prompt.
This feature is not available on the free tier. It becomes accessible starting from the Lite plan ($5.99/month), where it is restricted to short 3-second clips. Full-length video generation is available at Plus and above.
The video generator is a meaningful differentiator in the AI companion market. A direct survey of competitors:
- Character.AI: No video generation
- CrushOn AI: No video generation
- Janitor AI: No video generation
- Candy AI: Limited video capability
- Replika: No video generation
Among mainstream AI girlfriend platforms, Secrets AI is one of very few with this capability. Platforms with genuinely comparable video generation — such as SweetDream AI and Xotic AI (which offers 4K, 15-second clips) — represent the outer edge of the market rather than the mainstream competitors.
How Video Generation Works
The process has four steps:
- Select a source image — either from the character's auto-generated gallery (four images are created at character setup) or from any image you have generated in previous sessions
- Write a motion prompt — describe the movement, action, or expression you want to see (e.g., "walking toward the camera slowly," "laughing and tilting her head," "turning to look over her shoulder")
- Submit and wait — generation takes approximately two minutes
- View and save — the completed clip is available to view and download from the session
The video output is context-aware: it reflects the character's specific appearance from the source image, the environmental context of the original scene, and the motion described in the prompt. Complex multi-part prompts occasionally produce artifacts or miss elements of the description; simpler, specific prompts generally produce cleaner results.
Video quality improves on the advanced generation model available at Premium and Ultimate tiers. Lite and Plus users access the standard model; Premium and Ultimate access the advanced model.
Video Quality Assessment
Independent reviewers rate Secrets AI video quality at 4.1/5 — a strong score for this category, where video generation quality tends to be inconsistent across the market.
Reviewer observations that hold up in practice:
- Character movement is generally smooth, not robotic
- Facial expressions correspond to the motion prompt's emotional context
- Character appearance in video matches the source image accurately
- Occasional quality variations based on prompt complexity
- More complex scene descriptions produce higher artifact rates than simple motion prompts
The practical floor for quality is "good enough to be recognizable and natural-feeling most of the time." The practical ceiling — on advanced model, with a well-written prompt and a clean source image — is "genuinely impressive given that the character is AI-generated."
The known limitation shared with the image generation system is hand anatomy: close-up hand motions can produce the inconsistencies common across deep learning-based image and video generation.
How Much Do Videos Cost in Moments?
This is where video generation requires careful budgeting:
| Video Type | Moments Cost |
|---|---|
| Short clip (3 seconds) | ~50 Moments |
| Full-length clip | ~600 Moments |
For reference, a single full-length video clip costs as much as 12-24 images. This cost relationship is the primary reason video-heavy use patterns should be on Premium or Ultimate.
What each tier's Moments allocation buys in video terms:
| Tier | Monthly Moments | Full-Length Videos | Short Clips (3s) |
|---|---|---|---|
| Lite | 1,000 | ~1-2 | ~20 |
| Plus | 3,000 | ~5 | ~60 |
| Premium | 8,000 | ~13 | ~160 |
| Ultimate | 15,000 | ~25 | ~300 |
These figures assume Moments are spent only on video. Real mixed-use patterns (text + images + occasional video) produce different totals. A realistic Premium use pattern — 100 images, 5 full videos, 15 minutes of voice calls — consumes roughly 6,000-7,500 Moments, leaving some buffer.
Moments reset monthly with no rollover. Unused video capacity from one month does not carry to the next.
Bulk Moments top-ups are available starting at $5.99 for 1,980 Moments. For users who hit their monthly allocation before the cycle resets, top-ups are the bridge option before upgrading tiers.
Video vs Images vs Voice — Full Cost Comparison
| Feature | Moments Cost | Output Type | Cost Efficiency |
|---|---|---|---|
| Text message | 1-2 | Text response | Highest (cheapest/message) |
| Image generation | 25-50 | Static image | Good |
| Short video (3s) | ~50 | Motion clip | Similar to image |
| Full-length video | ~600 | Longer motion clip | Lowest (most expensive/output) |
| Voice call | 100/minute | Real-time audio | Depends on session length |
The ratio that clarifies the cost structure: 600 Moments buys one full-length video OR 12-24 images OR 6 minutes of voice calls. This comparison makes the video premium clear in concrete terms.
For Moments costs across all features, that page provides the complete pricing breakdown including subscription tier allocations and bulk top-up pricing.
Tips for Better Video Results
Practical recommendations based on how the generation system behaves:
- Use high-quality source images — video generation quality correlates with source image quality; blurry or low-detail source images produce lower-quality video output
- Write specific motion prompts — "walking slowly toward camera with a soft smile" outperforms "walking" because the specificity constrains the generation toward the intended result
- Start with a short clip before a full video — at 50 versus 600 Moments, a test clip costs 1/12th of a full video; verify the prompt direction before spending 600 Moments
- Use the advanced generation model — available at Premium and above; quality improvement is noticeable for video more than for images
- Generate the source image with video in mind — front-facing, good lighting, uncluttered background source images tend to produce cleaner video output than complex scene images
- Keep motion prompts simple and single-action — multi-part prompts ("walk toward camera, then turn and smile, then reach out") frequently produce artifacts; one clear action per clip is more reliable
Who Should Use the Video Generator?
Worth it if:
- Visual content is a core part of how you engage with the companion experience
- You want unique media assets from your AI companion rather than only static images
- You are on Plus or above — at Lite's 1,000 Moments/month, video budget is extremely tight
Not worth it if:
- You are on the free tier (video is unavailable)
- You are on Lite and primarily want conversation or images — 1,000 Moments allocated to video leaves very little for other features
- You are a text-first user — the Moments cost of video is high relative to the conversational value it adds
Best tier specifically for video use:
- Moderate video users (5-10 full clips/month plus images and text): Plus ($9.99)
- Regular video users (10-15 clips/month mixed with heavy image use): Premium ($19.99)
- Heavy video users (20+ clips/month): Ultimate ($39.99)
For video access by tier in the context of the full tier comparison, that page shows how video fits into the complete feature unlock table.
For the broader feature assessment including all generation capabilities, the full features page covers images, voice, and the complete memory system alongside video.
Access Secrets AI video generation →
And for the complete platform review with overall ratings and verdict, that page contextualizes video quality within the full platform assessment.
FAQ
Short clips are approximately 3 seconds, available from Lite tier at roughly 50 Moments each. Full-length clips are longer and cost approximately 600 Moments. The precise maximum clip length is not publicly documented as a specific number of seconds, but the "full-length" tier is described in reviewer assessments as meaningfully longer than the 3-second format.
No. Video generation is not available on the free tier. The minimum subscription required to generate any video is the Lite plan ($5.99/month), which provides access to short 3-second clips. Full video generation requires Plus ($9.99/month) or above.
This depends on your subscription tier and how you allocate Moments. On Plus (3,000 Moments/month), you can make approximately 5 full-length videos or up to 60 short 3-second clips if you spend all Moments on video exclusively. On Ultimate (15,000 Moments/month), approximately 25 full-length videos. Real mixed-use patterns produce lower numbers since Moments are shared across all features.
Video quality is rated 4.1/5 by independent reviewers. Character movement is generally smooth, facial expressions correspond to the motion prompt, and the character appearance matches the source image accurately. Quality improves on the advanced generation model available at Premium and above. Complex prompts and close-up hand motions have the highest artifact rate.