Secrets AI Video Generator: How It Works, Quality, and Cost
Video generation from AI companion images is one of the most genuinely distinctive capabilities in the AI companion market. Most competing platforms — Character.AI, CrushOn AI, Janitor AI — do not offer it at all. Candy AI offers limited video. Secrets AI has built it as a core feature, available from the Lite tier upward with a quality rating of 4.1/5 from independent reviewers.
This page covers exactly how video generation works on Secrets AI, what it costs in Moments, how to get the best results, and whether it justifies upgrading or spending additional Moments on.
What the Video Generator Actually Does
The Secrets AI video generator converts static AI companion images into short animated video clips. The input is an existing companion image plus a text prompt describing the desired movement or action. The output is a generated video clip of that companion performing the described motion.
This is not screen recording, not animation in the traditional sense, and not live video. It is AI-generated video produced from deep learning models (analogous to Stable Diffusion and related image-to-video architectures) that interpret visual content from the source image and render coherent motion based on the text prompt.
The practical implication: the quality of the output depends heavily on two factors — the quality and resolution of the source image, and the specificity and clarity of the motion prompt. Overly complex or contradictory prompts produce inconsistent results. Clear, specific prompts produce clips that closely match the intended action.
Availability: Video generation requires Lite tier or higher. The free plan has no video access.
Generation time: Approximately 2 minutes per clip, regardless of length. This is server-side processing — you wait for the clip to render.
The Generation Process, Step by Step
The workflow is consistent whether you are generating your first clip or your hundredth:
- Select or generate a source image — choose from the auto-generated character images created at character creation, or generate a new image (25-50 Moments). Higher-quality, well-lit images without background clutter produce better video output.
- Open the video generation option — available within the chat interface after selecting an image.
- Write a motion prompt — describe what you want the character to do. Specificity helps: "slowly turning head to the right and smiling" produces better results than "look at me." Action verbs, direction, speed, and facial expressions are useful parameters.
- Confirm and queue the generation — the Moments cost is deducted at this point. Short clips (~50 Moments), full clips (~600 Moments).
- Wait approximately 2 minutes — server processing time; you can continue other actions while waiting.
- Review and save — the completed clip appears in the interface. Save it to your device if desired.
Video Quality: Honest Assessment
Independent reviewers at aigirlfriendscout rate Secrets AI video quality at 4.1/5, described as videos that "look good and move smoothly most of the time." This is a fair characterization of what to expect:
What looks good: Character appearance from the source image is preserved accurately in the video. Facial expressions render naturally in most clips. Simple motions — head turns, smiles, gestures, subtle body movement — produce smooth, convincing results. The visual fidelity of the character carries through from image to video without degradation.
Where quality varies: Complex, multi-step motions produce less consistent results. Full-body movement is harder to render cleanly than upper-body or face-focused clips. Hands — already a documented weakness in AI image generation — remain an occasional problem in video clips with prominent hand movement. Longer clips (those approaching the higher Moments cost range) may show quality inconsistency between the beginning and end of the sequence.
The Premium generation model matters. Video generation quality is higher on the Advanced model available to Premium and Ultimate tier users. If video quality is a priority, the generation model tier is worth factoring into subscription decisions.
Moments Costs by Video Type
| Clip Type | Moments Cost | Monthly Budget by Tier |
|---|---|---|
| Short clip (3 seconds) | ~50 Moments | Lite: ~20 clips / Plus: ~60 clips / Premium: ~160 clips / Ultimate: ~300 clips |
| Full-length clip | ~600 Moments | Lite: ~1-2 clips / Plus: ~5 clips / Premium: ~13 clips / Ultimate: ~25 clips |
The cost asymmetry between short and full clips is the most important number to understand. A single full-length clip costs the equivalent of:
- 300-600 text messages
- 12-24 standard images
- 6 minutes of voice calls
For casual video use, the Plus tier (3,000 Moments) supports approximately 5 full-length clips per month alongside regular chat. For consistent video generation, Premium (8,000 Moments) provides approximately 13 full-length clips — a more practical budget. Heavy video creators on Ultimate (15,000 Moments) can generate approximately 25 full-length clips per month from their standard allocation.
Video vs Other Media: Cost Comparison
Understanding video costs requires context against the other media options:
| Feature | Moments Cost | What You Get |
|---|---|---|
| Text message | 1-2 | One AI response |
| Image generation | 25-50 | One static image |
| Short video (3 sec) | ~50 | Brief animated clip |
| Full video | ~600 | Longer animated clip |
| Voice call | 100/minute | Real-time audio response |
For the same 600 Moments it takes to generate one full-length video clip:
- 1 full video, OR
- 12-24 standard images, OR
- 6 minutes of voice calls, OR
- approximately 300-600 text messages
This comparison makes clear that video is a discretionary premium feature, not something you casually generate throughout a session. Users who value chat and occasionally want video should plan a monthly video budget (e.g., 2-3 clips) rather than treating video as a default action.
Moments top-ups are available if you exhaust your monthly allocation: 1,980 additional Moments cost $5.99, covering approximately 3 full-length video clips. For full details on Moments pricing by tier, see the subscription plans breakdown.
Tips for Better Video Output
Experience with the feature surfaces a set of consistent best practices:
Invest in image quality first. Video inherits from image — a sharper, more detailed source image produces a sharper, more detailed clip. Generating the highest-quality image available before attempting video conversion improves output noticeably. Use the Advanced generation model for source images if available on your tier.
Write specific, single-action prompts. "Smiling and slowly tilting head to the left" outperforms "be playful" in terms of output consistency. The AI interprets concrete motion descriptors more reliably than emotional or abstract instructions.
Start with short clips to test a prompt. At ~50 Moments per short clip versus ~600 for a full one, testing a prompt on a 3-second clip before committing to full length saves significant Moments if the initial output does not match expectations.
Avoid complex multi-step sequences. Prompts like "turn around, then wave, then smile" tend to produce less coherent results than single-action prompts. Generate multiple short clips for sequential actions rather than one complex long one.
Use full-length clips selectively. Reserve the 600-Moment cost for prompts and images you have already validated at short-clip quality.
Who Benefits from Video Generation
Genuine use cases: Users who want dynamic visual content from their AI companion, who enjoy saving or revisiting generated media, or for whom the visual dimension of the companion relationship is significant. Video brings a static character image to life in a way that additional images cannot replicate.
Less relevant for: Text-first users whose primary interaction is conversation — the Moments spent on video provide no conversational benefit. Users on tight Moments budgets who already find image and voice consuming their monthly allocation.
Tier recommendations for video users:
- Occasional video (1-5 clips/month): Plus ($9.99/month) — 3,000 Moments covers light video use alongside regular chat
- Regular video (10-15 clips/month): Premium ($19.99/month) — 8,000 Moments provides a sustainable video budget
- Heavy video generation: Ultimate ($39.99/month) — 15,000 Moments makes video a primary activity rather than a special-occasion feature
For a full breakdown of what each tier provides beyond video, including memory system depth and exclusive features, the free vs premium comparison covers all tier differences.
How Secrets AI Video Compares to Competitors
The video generation landscape in AI companions is sparse:
| Platform | Video Generation | Notes |
|---|---|---|
| Secrets AI | Yes (full feature) | 3-sec to full-length clips, Moments-based cost |
| Candy AI | Limited | Less developed than Secrets AI |
| CrushOn AI | No | Not available |
| Character.AI | No | Not available |
| Janitor AI | No | Not available |
| SweetDream AI | Yes | Comparable feature |
| Xotic AI | Yes | 4K, 15-second clips |
The competitive advantage of Secrets AI's video generation is real — most platforms that users compare it against do not offer the feature. Among platforms that do offer video (SweetDream AI, Xotic AI), Secrets AI's implementation is competitive on quality if not always on technical resolution.
For users who specifically want video generation as part of their AI companion experience, Secrets AI is currently one of the stronger choices by default of being one of the few choices.
Access Secrets AI video generation →
For the complete platform assessment, the full review covers all features, pricing, and an overall verdict.
FAQ
Video length depends on the clip type selected and your subscription tier. Short clips on Lite tier are approximately 3 seconds. Higher tiers and the full-length clip option produce longer videos — the exact maximum length is not publicly specified in seconds, but reviewers describe "longer motion clips" at the ~600 Moments tier. The generation time is approximately 2 minutes regardless of output length.
No. Video generation requires the Lite tier ($5.99/month) or higher. The free plan has no video access regardless of Moments balance. On Lite, the available clips are the shorter (3-second) format; full-length video generation becomes available from Plus ($9.99/month) upward.
This depends on your subscription tier and the Moments cost of your chosen clip type. At ~600 Moments per full-length clip: Lite (1,000 Moments) supports 1-2 clips; Plus (3,000 Moments) supports approximately 5; Premium (8,000 Moments) supports approximately 13; Ultimate (15,000 Moments) supports approximately 25. Short clips at ~50 Moments each produce higher quantities: Plus (3,000 Moments) supports approximately 60 short clips per month. Additional Moments can be purchased in top-up bundles starting at $5.99.
Independent reviewers rate the video quality at 4.1/5, describing the output as looking "good and smooth most of the time." Character appearance from the source image carries through accurately, and simple motions render convincingly. Quality varies with prompt complexity and generation model tier — Premium and Ultimate users access the Advanced generation model, which produces noticeably better output than the standard model. Occasional quality inconsistencies appear in complex motion sequences.