Melius

Node-based creative canvas where an agent named Mel plans and runs multi-model image, video, and audio generation workflows.

Community:

AI Photo & Image Generator AI Design Assistant AI Advertising Assistant AI Video Generator AI Graphic Design

Visit Website

Atoms - Build websites & apps easily with AI

StoreClaw

Sponsor

An autonomous commerce engine for online sellers that monitors, executes, and grows your store across all major e-commerce platforms — 24/7, without manual intervention.

Overview
Alternatives
Analytics

Atoms - Build websites & apps easily with AI

Product Overview

What is Melius?

Melius is a creative operating system built around an agentic canvas instead of a chat window. Users describe what they want in plain language, and Mel, the built-in agent, plans the workflow, picks the right model for each step, and lays it out as an editable graph of nodes and edges. Every prompt, model choice, and intermediate output stays visible on the canvas, so creators can open, tweak, and re-run a single node rather than redoing an entire pipeline. Melius routes work across leading generation models from providers like OpenAI, Google, ByteDance, Kuaishou, Ideogram, and ElevenLabs, and it can also be driven from outside the app through an MCP server that connects to Claude, ChatGPT, or Codex.

Key Features

Agentic Node Canvas
Work happens on an infinite canvas of nodes and edges rather than a chat thread, with every generation step exposed as something you can inspect, edit, and re-run.
Cross-Provider Model Routing
GPT Image 2, Nano Banana, Seedance, Kling, Ideogram, and ElevenLabs voices sit on the same workspace, so there's no need to juggle separate subscriptions or track which model is currently best.
Mel, the Planning Agent
Brief Mel in plain language and it writes the prompts, chooses models, and assembles the workflow, switching between auto-run and ask-permission modes depending on how much oversight you want.
MCP Server Access
The canvas is exposed over MCP at mcp.melius.com, letting Claude, ChatGPT, or Codex read the live node graph and build, wire, and run generations directly from an agent chat.
Real-Time Multiplayer
Human collaborators and AI agents appear on the same canvas simultaneously, so teams can steer a workflow together as it's being built.
Credit-Based Plans
Creator, Growth, Professional, and Enterprise tiers each allocate a monthly credit pool that scales seats, agent skills, and priority access as needs grow.

Use Cases

Agency Pitch Work : Agencies build concept boards and treatment decks at the brief's pace, then scale the same canvas into campaign variants and spec ads once the pitch is won.
Filmmaking and Storyboarding : Directors and filmmakers turn a described shot into storyboards, AI shorts, and reference boards using tunable, multimodal nodes instead of a single-model tool.
Localized Campaign Variants : Marketers generate thousand-variant ad cascades, animated statics, and localized copy for every market and channel from one hero shot.
E-commerce Product Imagery : Teams produce pack shots, on-model imagery, and PDP variants that stay brand-consistent, cutting a multi-week shoot down to an afternoon on the canvas.
GTM and Growth Assets : Growth teams skip the design queue and build event graphics, blog heroes, and sales deck visuals directly, without learning a tool or writing a prompt.

FAQs

StoreClaw

Sponsor

An autonomous commerce engine for online sellers that monitors, executes, and grows your store across all major e-commerce platforms — 24/7, without manual intervention.