PricingSave big

How to use the Picsart CLI for AI image, video, and audio generation

SKILLS8 minIntermediate

Master the complete gen-ai CLI workflow from installation to batch generation across 130+ frontier AI models.

How to use the Picsart CLI for AI image, video, and audio generation

What you'll learn

  • Install and authenticate the Picsart CLI on any platform
  • Generate images, video, and audio with a single command
  • Pick the right AI model for your creative task
  • Run batch generations and pipe outputs to other tools

What is the Picsart CLI?

The Picsart CLI (gen-ai) is a terminal tool that connects you to 130+ frontier AI models for image, video, and audio generation. Also known as the Picsart gen-ai CLI or Picsart MCP server, it brings models like Flux, Sora, Kling, Veo, Runway, Recraft, and ElevenLabs into one install. Think of it as a universal remote for creative AI — instead of juggling five different vendor accounts, you type one command and get your file.

Common use cases

  • Content creation: Generate hero images, social graphics, and video clips for campaigns
  • Prototyping: Mock up visual concepts before hiring a designer or photographer
  • Batch workflows: Re-render entire folders of product images with new backgrounds
  • Agent tooling: Give Claude Code, Cursor, or Windsurf native creative generation abilities
  • CI/CD pipelines: Auto-generate OG images during Next.js builds or cron jobs
  • Podcast production: Turn scripts into voiceovers with ElevenLabs voices in seconds

Use the Picsart CLI step by step

STEP 1: Install the CLI

  • On web: Go to picsart.com/cli → Copy the macOS/Linux/Windows install command
  • On mobile: Not applicable — CLI runs in terminal environments only
Get install command

STEP 2: Authenticate your account

Run gen-ai login to connect your Picsart account:

  • A browser tab opens automatically
  • Click "Allow" to authorize the CLI
  • Credentials are stored locally on your machine
  • No API keys to copy or paste — authentication happens once

STEP 3: Generate your first file

Run gen-ai generate with a prompt and optional model flag. The CLI picks a smart default model if you don't specify one. Your file saves to the current directory automatically.

STEP 4: Review and iterate

Check that the output matches your intent: Not perfect? Run the command again for a new variation, switch models with -m, or tweak your prompt for better results.

  • Look for visual quality and composition accuracy
  • Verify aspect ratio and resolution match your needs
  • Check that the style aligns with your brand or project
Start generating

Tips for best results

💡 Start with the default model

The CLI auto-selects a good model for your prompt type. Once you see the result, switch models if you need a different style or speed. Recraft for vector-style graphics, Flux for photorealism, Sora for cinematic video.

💡 Use --json for scripting

Add --json to any command to get structured output you can pipe to jq, curl, or CI tools. Perfect for build scripts, batch jobs, or chaining multiple generations together.

💡 Check credit cost before running

Run gen-ai estimate with your prompt to see the credit cost upfront. Video and long audio clips cost more than single images. Batch jobs show total cost before you commit.

💡 Wire skills into your AI agent

If you use Claude Code, Cursor, Windsurf, or ChatGPT, load the gen-ai-use skill so your agent can generate media without you typing commands. Just say what you need in plain English.

Command quick reference

  • gen-ai generate: Create an image, video, or audio file from a text prompt
  • gen-ai login: Authenticate your Picsart account via browser
  • gen-ai models: List all available AI models and their capabilities
  • gen-ai estimate: Calculate credit cost before running a generation
  • gen-ai batch: Process multiple prompts from a JSON manifest or directory
  • gen-ai extend: Chain video extensions to create longer sequences
  • gen-ai drive sync: Upload outputs to Google Drive or Dropbox automatically
  • gen-ai --help: Show full command documentation and examples

Frequently asked questions

The CLI auto-selects a model based on your prompt, but you can override with -m. Use Flux or Recraft for images (Flux for photorealism, Recraft for vector-style graphics), Sora or Kling for video (Sora for cinematic quality, Kling for faster turnaround), and ElevenLabs for voiceovers. Run gen-ai models to see the full list with descriptions and credit costs.

No local install is required if you use the web playground at picsart.com/ai-playground. You can test prompts and models in your browser, then download results directly. The playground uses the same backend as the CLI, so results are identical. For automation, batch jobs, or agent integration, install the CLI.

The CLI itself is free to install and use. Generations consume credits tied to your Picsart account. Free accounts get a starter credit balance. Each model has a different credit cost — images start at 1-2 credits, video clips cost more depending on length and resolution. Check pricing at picsart.com/cli or run gen-ai estimate before generating.

Video models (Sora, Kling, Runway) take 30 seconds to several minutes depending on length and resolution. Image models finish in 5-15 seconds. Audio generations are fast (under 10 seconds for most voiceovers). If a job stalls, the CLI shows progress updates. You can cancel with Ctrl+C and try a different model or lower resolution.

Ready to start generating?

Install the CLI and connect 130+ frontier AI models to your terminal in under a minute.

Get the CLI