Template
- Campaign
- SKU
- Market
TEMPLATES
Build it once. Run it forever.
Turn your best workflow into a template, then apply it programmatically to every new campaign, SKU, or market. The next launch starts 80% built.
API
Run any generative AI model. One API.
The latest image, video, audio, and language models, all behind a single key. Generate one asset or build an entire pipeline.
Every model. No extra integrations.
GPT, Claude, Gemini, Flux, Nano Banana, Veo, Sora, Kling, Seedance, ElevenLabs, and more. One key, one API, no provider accounts to juggle. When a new model ships, it's already there. Choose a model per request, or let Auto pick the right one for the job.
GPT
Claude
Gemini
Flux
Nano Banana
Veo
Sora
Kling
Seedance
ElevenLabs
- GPT-6 Astra
- Nano Banana Pro
- Veo 3.1
- Kling 3.0 Omni
Auto
- Claude Opus 5.5
- Flux 2 Max
- ElevenLabs v3
- Seedance 2.5
- Gemini 3.1 Pro
- Seedream 5.0 Pro
- GPT-6 AstraText
- Nano Banana ProImage
- Veo 3.1Video
- Kling 3.0 OmniVideo
- Claude Opus 5.5Text
- Flux 2 MaxImage
Auto
- ElevenLabs v3Audio
- Seedance 2.5Video
- Gemini 3.1 ProText
- Seedream 5.0 ProImage
MiniMax Music 3Audio
MODELS
Model: Auto
Choose a model per request, or let Auto pick.
Script
Storyboard
Finished ad
STEP BY STEP
Chain models together.
Wire a text model into an image model into a video model, then run the whole chain in one call. Script to storyboard to finished ad, each step feeding the next.
FORMATS
Images, video, audio, and text
Built-in tools like background removal and resizing
Make one image, or a whole campaign
For marketers
Test more ads
Got an ad that's working? Make 50 versions and see which one wins.
For global brands
Launch in every market
Translate one campaign for every country
For social teams
Fit every screen
Turn one finished image into square, vertical, and widescreen
For e-commerce
Skip the studio
Put your product on a clean background or out in the real world
Plug Melius into anywhere you work
Connect Melius to Claude, ChatGPT, Codex, or any agent with our MCP server. Brief your agent in plain language, and it builds canvases, runs generations, and pulls from your asset library for you. Prefer the terminal? The mel CLI does the same from any script.
Set up MCP