Continue
·
Start
Describe an image or a video and every model is one picker away.
Your last prompt will wait here next time.
Credits
Unlimited
Loading…
Create
Recent work
Nothing here yet
Your images and videos appear here as you make them.
Best for:
Describe what you want to see in the image. Be specific about subjects, setting, lighting, and mood. Click "Enhance" to let AI improve your prompt.
No history yet
Tip: Use @image1, @image2 etc. in your prompt to reference specific images
Upload multiple reference images. Use @image1, @image2, etc. in your prompt to tell the AI how to use each one. Example: "Recreate @image1 with the logo from @image2".
Upload product photos, logos, labels, etc.
Flux supports only 1 reference image, the first image will be used
Pick a look to instantly set camera, lens, film, and lighting. Or customize each setting individually below.
This camera has a fixed/built-in lens
Seeds make images reproducible. Same seed + same prompt = same image. Leave empty for random. Only works with Flux model.
Seed only available for Flux
Drop a picture whose composition you want to keep: where the bottle stands, where the model poses, the horizon. The prompt describes the new picture.
Upload a layout image
Its composition is kept, the prompt fills it in
Layout image attached
Renders a 1024 square. Flux takes one image, so references and image to image are off.
Upload a source image and describe changes. The AI will edit/transform it based on your prompt. Not available for Seedream.
Not available for Seedream
Upload source image
The AI will transform this image
Source image attached
Ready for img2img
Low = subtle changes, keep source details. High = major changes, more creative freedom.
Gemini handles variation naturally via prompt, no strength slider needed
Tip: Be specific about what to preserve, the AI will only change what you describe.
The scene, pose, and camera angle stay the same, only the person changes.
Upload an image whose visual style you want to copy. The AI will generate new content in that style.
Add style reference
Copy the visual style of another image
Style reference attached
Style will be transferred
Ad Variations: Upload a competitor ad or existing creative. AI analyzes its style, colors, and audience, then generates variations you can customize.
Drop reference ad
PNG, JPG, WEBP up to 10MB
Reads the ad's style, colors, layout and audience so the variations keep what works.
Colors
Layout
Audience
Style
Upload 2-30 images. The AI will cluster them by visual style and extract reusable patterns from each cluster.
Click to upload multiple images
2-30 reference images for bulk analysis
Clusters Found
1:1 for Instagram/Facebook feed. 16:9 for YouTube/website banners. 9:16 for Stories/Reels/TikTok. 4:5 for Instagram portrait. 21:9 for cinematic/ultrawide.
Fastest, about 10 seconds per image.
Saver applies to Generate. Branch, Grid and Smart Shot run on Max.
Drop a product image (or click to browse)
Same source feeds every selected shot type
Source ready
Each selected shot type will use this image
Last batch, shots
Each shot is also saved to your gallery with a `smart-shot` tag.
Upload an image, then describe edits in chat. Each message edits the latest version (multi-turn). Powered by Nano Banana 2.
Drop an image (or click to browse)
jpg, png, webp · 10MB max
Current version
Tip: click any assistant image above to revert and branch from that version. All results auto-save to your gallery.
Open this on your computer
The overlay editor works best with a mouse. Open Viral Engine on your laptop or desktop to design posts. Everything else (image / video generation) works on mobile.
Drag any layer to move it. Double-click text to edit. Pick a template on the left to start fast.
Tap a logo to drop it on the canvas.
Pick a template above to start fast, or click Add Text / Shape / Logo to build from scratch.
Click a layer to edit it. Drag to move. Double-click text to type.
Save as template
Name this template so you can find it later. It saves to your active agent () and shows up in "Your saved templates" above.
Dramatic reinterpretations, different angles, environments, and compositions
Near-identical, only minor variations in details and subtle color shifts
Click to use as source image:
Best for:
Drag and drop, click to upload, or pick from gallery below
Source image ready
Or pick from recent generations:
Marketing Studio is off, Seedance will run on your raw prompt below, no presets or scaffolding.
Final prompt preview
Higgsfield public API caps at 12s. Cost: credits (60 per second at 720p, 120 at 1080p).
Describe items you can reference in your prompt by name. Type @product, @avatar, or @backdrop in the prompt and they'll be replaced with these descriptions on submit.
Tip: pick a style, hook, and setting, then add product details to the prompt below to ground the generation.
Drop / pick an end frame to lock the final shot
End frame ready
Video will land on this image
Upload the character image to animate
Character image ready
Motion will be applied to this character
Upload a video of the motion to mimic (.mp4/.mov, max 100MB)
3-30 seconds, best with clear human movement
Motion video ready
The image above is the reference. The result runs as long as your video: 4 to seconds. Object swap needs a video of at least 640x640.
Match Video: character faces same direction as in the motion video (best for complex movements). Match Image: character keeps the orientation from the uploaded image.
Upload character image
Character ready
Upload second character
Character 2 ready
Select a template to populate the prompt. Replace [YOUR MESSAGE HERE] with your content.
1:1 renders on Kling, Luma and Seedance; other engines output 16:9 or 9:16.
Select at least one output format
Generating...
Generating Video
Generation continues even if you leave
Generation failed
Could not price this resume.
Video Expired
Temporary URL no longer available
URL expired, regenerate to get a new video
Extend Video
Each agent bundles a master prompt, style, model and brand settings, and can be reused anywhere you generate.
No agents yet
Create an agent to save generation defaults for a client or project.
Your Funnels
Create high-converting landing pages and capture leads
No funnels yet
Create your first funnel to start capturing leads
All Leads
total leads
| Contact | Status | Source | Date | Actions |
|---|---|---|---|---|
|
|
||||
|
|
||||
|
No leads yet Leads will appear here when people submit your forms |
||||
Email Sequences
Automated follow-up campaigns
SMS Campaigns
Text message automation
No SMS campaigns yet
Webhooks
Connect to external services
No webhooks configured
Integrations
Connect your favorite tools
Total Visits
+24%12,847
Last 30 days
Avg. Session
+12%2m 34s
Engagement time
Bounce Rate
+3%42.3%
Single page visits
Performance Overview
Analytics chart will appear here
Connect analytics to see data
No VSL projects yet
Create one to start building your script
AI VSL Machine
4-stage system for creating VSLs that convert.
Research, Script, Produce, QA, powered by direct response principles from Schwartz, Hopkins, Bencivenga, Halbert, and Kennedy.
Build a VSL video writes the script, plans every scene, voices it, animates the key moments with Higgsfield and hands you a finished video.
Start from an example ?
Or skip this and describe your own product below.
Your brief
words, about min of narration
Length ?
Style ?
Shape ?
Captions ?
Voice ?
Motion with Higgsfield: ?
More: product photos, brand, music
Product photo links ?
Brand agent ?
Music link ?
Estimated cost ?
The price could not be loaded.
Nothing is rendered yet. You review the script and every scene first.
This takes about a minute for every 3 minutes of video. You can leave this page; it keeps going.
Planning stopped: . Your credits for it were refunded.
Script ?
scenes. Set motion for: ?
Exact price ?
About of video.
Start the render for ? It runs in the background and takes a few minutes per minute of video.
You can close this page. The render keeps going and the finished video lands here and in your gallery.
The render stopped: . Your credits for it were refunded in full.
This prompt requires no inputs, it teaches the AI the arguments-first framework.
Output
Previously Generated
Paste research from Perplexity, ChatGPT, or other sources. This will be saved to your project for use in prompts.
How Voice Changer works
Upload an MP3 of yourself reading the script with the right delivery, then pick a target voice. ElevenLabs preserves your performance, pacing, emphasis, breath, and re-renders it in the new voice. Costs 50 credits per minute (rounded up). Max 5 minutes per upload.
Drop an audio file here, or click to browse
MP3, WAV, M4A, or OGG · Up to 25MB · Up to 5 minutes
·
Lower = more expressive, higher = more consistent
How closely to match the target voice's timbre
Recording length , billed per started minute.
Converted
Conversion failed
Suno writes the lyrics for you.
Music is temporarily unavailable. Try again later.
You get takes of every song. You are only charged if they are made.
, ,
Cutting your edit to the beat. This usually takes a few minutes.
This song has a loose beat, so some cuts may feel less locked in.
Your songs will show up here
Describe one on the left to make your first.
Take ·
Edit a video to this song
Footage ( chosen)
Upload a video, or make one in Video Factory first.
Style
Look
Captions (we listen for the words, so check them)
Shape
Length
Song tools
·
Using the song you uploaded. Extend, replace and WAV need a song made here.
This song was made before these tools, so only cover, vocals and backing work on it.
At least 10 seconds, at most half the song.
Make a music video
Your product or brand images (up to , each clip starts from one)
Engine
Style
Shape
Length
Captions
clips of seconds at credits each, plus for the edit. You only pay for what gets delivered.
Timeline
shots, s, . Every cut sits on a beat.
Playing your last version. Make a preview to see your changes.
Style (locked shots stay)
Look
Captions
Shot , s, action on the beat, clear product shot, has text on screen
Waveform unavailable for this song. The beats are still marked.
Drag a white line to move a cut. It snaps to the beat. Arrow keys move it one beat.
Product images (optional, for the video later)
Reading what people say
. You can leave this page and come back.
Choose an angle
Evidence can come from similar products when yours has few public reviews. Each quote says whose it is.
Source From what you pasted
Make it
Image ads
Hooks
Scripts
Test
Name them
What to look at
When to stop
Ads are not launched from here: this is the plan for your ad account.
Earlier campaigns
All Models Included
No API keys needed. Your plan includes access to all AI models. Choose your preferred model when creating.
Interface
Legacy tools in the sidebar
Funnels, Leads, Automation and Analytics. Off by default; nothing is deleted either way.
Available Models
All included with your credits
Image Generation
Video Generation
Credit Costs
How credits are used per action
Command Line Access
Generate images and videos from your terminal, or hand it to your AI (Claude, Cursor, Copilot, etc.)
1. Install
npm install -g viral-engine-cli
Requires Node.js 18 or newer.
2. Log in from the CLI
ve login
Opens your browser to authorize this machine. Headless server? Use ve login --password.
3. Hand it to your AI
You have access to the Viral Engine CLI, an AI creative platform for image and video generation.
Setup (one-time):
npm install -g viral-engine-cli
ve login (opens browser; user approves this machine)
ve login --password (headless fallback if no browser available)
Discover commands:
ve --help
ve images --help
ve videos --help
ve gallery --help
ve videos providers (live list of video providers with credit costs)
ve images models (live list of image models with credit costs)
Common flows:
ve images generate "your prompt" --model nano-banana-2 --output ./out
ve images generate "your prompt" --model flux --output ./out
ve videos generate "your prompt" --provider sora -d 8 --output ./out --watch
ve videos generate "your prompt" --provider veo -d 8 --output ./out --watch
ve videos animate --image ./photo.jpg -p "gentle zoom" --provider kling --watch --output ./out
ve videos chain --script @script.txt --provider veo -d 30 --watch --output ./out
ve gallery list --favorites
ve credits balance
Global flags:
--json machine-readable output (use this for scripting)
--debug print requests + responses
-o, --output download directory
-w, --watch poll job to completion before returning
Notes:
- Every generation deducts credits (check with ve credits balance).
- Sora and Veo reject real-people / celebrity likeness. If they do, retry with --provider kling-3.0.
- Video jobs are async. Use --watch to block until done, or check with ve videos job <id>.
- The CLI ships an MCP server too: run viral-engine-mcp to use it with Claude Desktop.
Updating:
- The MCP server auto-reports if it is out of date on npm.
- When it does, tell the user to run this in their terminal:
npm install -g viral-engine-cli@latest
then fully quit and reopen Claude Desktop so a fresh MCP process spawns.
- npm install is not something Claude Desktop can run itself; it must happen in a real terminal.
- You can call the check_cli_version MCP tool at any time to re-verify version state.
Docs: https://viralengine.art/cli
Paste this block into your AI's context (Claude, Cursor, Copilot, etc.) so it knows how to use the CLI.
4. MCP for Claude Desktop
{
"mcpServers": {
"viral-engine": {
"command": "viral-engine-mcp"
}
}
}
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or the equivalent on Windows / Linux, then restart Claude Desktop.
5. Updating
npm install -g viral-engine-cli@latest
Updates need a real terminal. Claude Desktop cannot run npm on its own. After the install completes, fully quit and reopen Claude Desktop so a fresh MCP process is spawned. The MCP server also announces itself as out of date to the LLM, so your AI will prompt you to run the command above when it sees a new version on npm.
Start from a template
My Workflows
Pick a template above to get started, or create a blank workflow.
Drag nodes to canvas
Drag to pan canvas
Scroll to zoom
Drag ports to connect
Close or collapse the AI helper panel and the board comes back. Your board is still open.
Creative Studio is a drag-and-drop canvas. Open it on a laptop or desktop; everything else in Viral Engine works here on your phone.
No boards yet
Create one, then drop gallery images, notes and frames on it.
items, updated
Every account and model: what was made, what it was charged, what it cost us.
·
The report could not be loaded
Cost by model
Bar: provider cost
All accounts. Margin is credit value minus cost.
How to read this
What each number is, and what it cannot see.
- Exact
- Kie's own charge for the render, delivered or not.
- Est.
- The provider's list price for that model, size or length.
- Mixed
- More than one of these in one row, e.g. some renders ran through Kie.
- Unknown
- No price we can stand behind (VSL builds, scene effects); left out of the cost.
Accounts
| Account | ||||||
|---|---|---|---|---|---|---|
|
Unlimited
|
+? | |||||
|
|
||||||
No accounts match
Try another email or plan filter.