Picsart CLI & MCPAI generation for your terminal & agents

One catalog, three interfaces ​

These docs cover the developer & agent surfaces for Picsart's AI catalog. You can drive the same 176 models three ways:

• gen-ai CLI — a terminal app for the entire model catalog. Great for creators scripting their workflow and developers automating pipelines.
• Skills — drop-in .zip bundles that let you generate in plain English inside Claude Code, Cursor, Windsurf, or ChatGPT.
• MCP — the catalog exposed via Model Context Protocol, so any MCP-compatible agent can generate media directly.

All three share the same model registry and drive the same gen-ai engine — install the CLI, sign in once with gen-ai login (OAuth web login), and every surface works.

Prefer a visual UI? The AI Playground app ↗ is the web app — the point-and-click version of this same catalog, where you pick a model and generate in the browser. These docs are for using that catalog from the terminal and AI agents instead.

CLI quickstart ​

npm install -g @picsart/gen-ai      # or: curl -fsSL https://picsart.com/gen-ai-cli/install.sh | bash
gen-ai login                        # OAuth web login (opens your browser)

# text to video
gen-ai generate -m seedance-2.0 -p "a fox running through autumn leaves" -d 8
# image
gen-ai generate -m flux-2-pro -p "studio shot of a ceramic cup" --ar 4:3

MCP quickstart ​

Connect Picsart to your agent, then generate with a tool call:

// the same image generation, as an agent tool call
{ "name": "picsart_generate",
  "arguments": { "model": "flux-2-pro", "prompt": "studio shot of a ceramic cup", "aspectRatio": "4:3" } }

Full details in the MCP Quickstart. Browse everything in the Model Catalog or by Provider.