How to generate a video on Secrets AI
Generate a video on Secrets AI, screenshot by screenshot. One video costs 11 images: the method to test your prompt on the cheap before paying full price.
Demo — the starting image, and what it becomes once animated:
- You need an image first — the video animates from a first frame, not from text alone
- Advanced mode turns on audio and top quality by default — that is what drives the bill up
- ⚠️ A video costs 540 moments minimum, that is 11 images. Test your prompt on Fast + 480p + 5 s before moving up
- 4 to 15 seconds of duration, and the cost follows the slider
For the full verdict on the app, see our Secrets AI review.
This is the guide where I mostly save you from wasting money. Video on Secrets is good — but it costs eleven times an image, and the app defaults you to the most expensive settings.
The trap is simple: you write a rough prompt, launch at maximum quality, dislike the result, and you have just burned 540 moments for nothing. Do it twice and that is an eighth of your month.
How do you generate a video on Secrets AI?
To generate a video on Secrets AI: open the Generation Studio, switch to the Video tab, pick Advanced mode, load an image into First frame, describe the motion you want, then set Variant, Resolution and Duration before launching. Budget 540 moments at the cheapest setting.
The walkthrough, and the method to avoid waste, below.
Before you start
- A Secrets AI account with at least 540 moments
- A starting image. The video animates from a first frame — if you do not have one, start by generating an image, that costs 50 moments
- Some patience: rendering takes longer than for an image
Step 1 — Advanced mode, and what it implies
In the Generation Studio, Video tab, you first pick a Content Type. Three options: Adult, Advanced and Standard (flagged Legacy).

Advanced is the right pick for quality — it advertises the best prompt adherence, the highest quality, 4 to 15 seconds, and both Quality and Fast variants.
But read the third line: “Audio by default”. Advanced mode switches on audio and high quality automatically. It is comfortable, and it is exactly what blows up your bill while you are still hunting for the right prompt.
Step 2 — Load your first frame
Switch to the Video tab at the top right of the studio. Under Frames you get two slots: First frame (required) and End frame (optional).

Load your image into the first one. The description field below is not for describing the scene — that already exists — but for describing the motion: “walking slowly forward, hair moving gently, slight smile”.
Worth knowing: End frame is optional but powerful. Giving it a closing image constrains the motion path instead of letting the engine improvise.
Step 3 — The method that stops the waste
Here is the heart of this guide. Before launching anything, scroll down to Video Settings.

Three settings drive the price:
- Variant — Quality (up to 4K) or Fast (up to 720p, quicker output)
- Resolution — 480p, 720p, with 1080p and 4K unlocking in Quality mode
- Duration — from 4 to 15 seconds
My method, and the one I would recommend:
- Switch to Fast, 480p, 5 seconds. The cost drops to 540 moments, the floor
- Launch, and check whether the motion works. Not the image finesse — the motion
- If it missed, fix the description and relaunch at low settings. You lose 540 moments, not 1,500
- Once the motion is validated, only then move up to Quality and higher resolution
You pay full price once, on a prompt you already know works. Across three attempts, the difference runs into thousands of moments.
The scale to keep in mind: an image at 1.5K costs 50 moments. A video at the minimum costs 540. One video is eleven images. You do not test a video prompt the way you test an image prompt.
Step 4 — Launch and wait
The button shows the cost before you confirm. Click, and generation starts — it appears in the right-hand gallery with your prompt as a caption.

It takes longer than an image. You can leave the screen, rendering continues in the background and waits for you in the Videos tab of the gallery.
Frequently asked questions
How much does a video cost on Secrets AI?
540 moments at the cheapest setting — Fast variant, 480p, 5 seconds. The cost rises with resolution and duration, and the button always shows the total before you confirm. For scale: with Premium’s 8,000 monthly moments, that is roughly 14 videos at the minimum, against 160 images.
Can you generate a video without a starting image?
No. Secrets works image-to-video: you need a first frame. Starting from scratch means generating an image at 50 moments, then animating it. That is also what keeps the result coherent — the character does not change face halfway through.
Should you turn off audio to save moments?
Audio comes with Advanced mode, which enables it by default alongside high quality. In the testing phase, what you are after is validating a motion, not listening to a soundtrack. Switching to Fast brings you back to the floor price — you will move up, audio included, once the prompt is right.
What is the End frame for?
To constrain where the motion ends. With a single image, the engine decides on its own where it takes the scene. Giving it a closing image defines the path. It is optional, but it is the setting that gives you the most control over the result.
How long can a video be?
From 4 to 15 seconds, set with a slider. The app recommends 5 seconds. Longer costs more — and over long durations, the risk of motion drifting increases.
Going further
— Generate an image on Secrets AI — the step before, 50 moments — Our full Secrets AI review — pricing, moments, what the app is actually worth — How to cancel Secrets AI — and why you should download your creations first — Animate a photo with Promptchan — to compare the two video approaches — Best AI Girlfriend apps 2026
Generate your first video on Secrets →