Connect Picsart to your agent — and generate images, video, and audio right from your conversations.
Connect Picsart to your CLI or AI agent - powered by Nano, Banana, Flux, Sora, Kling, Veo, Runway, Recraft, ElevenLabs, and 120+ more models. Drop it in any terminal, CI job, or Dockerfile. MCP server for Claude Code, Cursor, and ChatGPT is live now.
Connect Picsart to your agent — and generate images, video, and audio right from your conversations.
One command installs the gen-ai binary on macOS, Linux, and Windows — no Node.js required.
curl -fsSL https://picsart.com/gen-ai-cli/install.sh | bash
OAuth opens in your browser. Credentials stay on your machine and unlock 130+ models on one credit balance.
gen-ai login
Describe what you want in plain English. Switch providers with -m. Output lands in ./out/.
gen-ai generate -m flux-2-pro -p "editorial hero" --ar 5:3
Choose the agent you use, then open its connector settings.
Name it Picsart and paste the URL below. Click Add.
https://api.picsart.com/gen-ai/mcp
Click Connect, sign in with your Picsart account, and ask your agent to generate an image.
One command pulls all the Picsart skills — generate, batch, brand-gate — into your agent.
npx skills add PicsArt/gen-ai-skills
Connects your account so the skills can submit jobs through the gen-ai CLI under the hood.
gen-ai login
Works with Claude Code, Cursor, Codex, Windsurf, and 12+ agents. Skills auto-trigger on plain-English prompts.
/picsart:generate "launch kit for product.md"
One CLI for images, video, and audio powered by Nano Banana, Flux, Sora, Kling, Veo, Runway, Recraft, ElevenLabs, and 120+ more leading models. Skip the SDK sprawl: one install, one auth, every provider. Drop it into any terminal, CI job, or Dockerfile to script creative workflows, batch-generate assets, or wire generation straight into your product. Same engine that powers Picsart's own agents, now open to every builder. MCP server for Claude Code, Cursor, and ChatGPT is live now.
One MCP server for images, video, and audio, powered by Nano Banana, Flux, Sora, Kling, Veo, Runway, Recraft, ElevenLabs, and 120+ more models. Install it once, connect your host, and ask in plain English. Claude Code, Codex, Cursor, ChatGPT, Windsurf, and other MCP-compatible tools can call Picsart generation without leaving your agent. The CLI is still available when you want raw shell control.
One CLI for images, video, and audio powered by Nano Banana, Flux, Sora, Kling, Veo, Runway, Recraft, ElevenLabs, and 120+ more leading models. Skip the SDK sprawl: one install, one auth, every provider. Drop it into any terminal, CI job, or Dockerfile to script creative workflows, batch-generate assets, or wire generation straight into your product. Same engine that powers Picsart's own agents, now open to every builder. MCP server for Claude Code, Cursor, and ChatGPT is live now.
Plain English in. Real creative output out. You don't need to learn a single command. Your AI agent speaks CLI so you don't have to. You bring the imagination, the taste, the idea. It handles the syntax.
No tricks. No decade of computer science. A command-line interface is a text box where you type an instruction and the computer writes back. The scary-looking window is actually the simplest kind of software there is.
You don't have to memorize the commands. Tools like Claude Code and Cursor read your plain English and write the CLI for you. Drop a gen-ai skill into your agent once, then trigger it with sentences. You bring the taste. The agent handles the syntax.
The barrier to entry has never been lower.
The value of taste and imagination has never been higher.
Agents change who gets to build. We are all builders. Or can be. What you bring is the thing no tool can manufacture: your eye, your voice, your weird idea at 2am.
Plain English in. Real creative output out. MCP, short for model context protocol, lets your AI agent call approved tools directly. Picsart MCP gives it creative tools for images, video, audio, model info, pricing, and estimates.
No app switching. No copied API payloads. Your agent gets a clear list of Picsart actions it can call, then chooses the right one based on what you ask for.
You do not need to memorize flags or model settings. Ask for the outcome, review the estimate when credits are involved, and let the agent call the right Picsart tool.
Your agent gets the production tools.
You keep the taste and direction.
Use MCP when you want creative generation inside chat. Use the CLI when you want terminal control, scripts, or batch jobs. Both point back to the same Picsart account.
Plain English in. Real creative output out. A skill is a drop-in folder that teaches your AI agent a complete job — generate a hero, cut a reel, re-shoot a catalog. Install one, ask for the outcome, and the agent runs it end-to-end.
No prompt engineering, no flags to remember. Each skill packages the instructions and the right Picsart tools, so a sentence is enough to trigger the whole workflow.
Drop a skill into your agent once. When your request matches — "build a launch kit for this product" — the skill fires, picks the model, and delivers the assets where you can use them.
Your agent gets the playbook.
You keep the taste and direction.
Skills work in Claude Code, Cursor, Codex, Windsurf, and 12+ agents. Under the hood they use the same Picsart account and credits as the CLI and MCP.
Run this in your shell — works in any terminal, CI job, or Dockerfile.
curl -fsSL https://picsart.com/gen-ai-cli/install.sh | bash
iwr -useb https://picsart.com/gen-ai-cli/install.ps1 | iex
npm install -g @picsart/gen-ai
First run opens your browser to sign in · Or use without installing →
Install the MCP server, register it with your host, then verify the connection.
Cursor, Windsurf, and VS Code can use gen-ai-mcp as the MCP command ·
Open docs →
Add the skills to your agent, sign in once, then ask in plain English.
npx skills add PicsArt/gen-ai-skills
gen-ai login
/picsart:generate "launch kit"
Works in Claude Code, Cursor, Codex, Windsurf, and 12+ agents · Browse the skill library →
gen-ai generate -m recraftv4 -p "editorial still-life, neon rim-light" --ar 1:1
"Render a 16:9 editorial still-life with neon magenta and cyan rim-light — use Recraft V4, save as hero.webp."
gen-ai generate -m kling-v3-standard -p "neon liquid ribbons" --duration 5 --ar 16:9
"Animate a 5-second neon-liquid ribbon loop in 16:9 — use Kling 3.0 Standard. Keep it cinematic, slow motion."
gen-ai generate -m eleven-v3 -p "Welcome to Picsart. One CLI for every media type."
"Voice this line in a warm, confident read with Eleven v3: 'Welcome to Picsart. One CLI for every media type.'"
One CLI. Three media types. Real output, generated with gen-ai.
Copy any prompt below and drop it into Claude Code, Cursor, or Windsurf. The installed skill picks the right model and runs it — no flags, no config.
"Shoot a dramatic editorial portrait of a woman in a rain-soaked Tokyo alley, neon signs reflected in puddles, 35mm film look."
"Golden hour landscape of an abandoned greenhouse overgrown with wildflowers — cinematic color grade, shallow depth of field, 16:9."
"Moody still life of vintage camera equipment on aged oak — single-source window light, heavy film grain, 4:3."
"A giant open book floating above a misty mountain range at dusk — dramatic volumetric lighting, hyperrealistic, wide."
"Black-and-white portrait of a jazz musician backlit in a smoky 1950s club — high contrast, Kodak Tri-X grain, 3:4."
"Editorial fashion shot — model in a tailored charcoal coat in an empty brutalist parking garage, harsh overhead strip light, 2:3."
"Hand-painted botanical illustration of a rare orchid species with scientific annotation labels — watercolor on textured paper, 3:4."
"Retro 1970s road trip poster — Route 66, desert sunset, bold two-tone color palette, vintage typography, 2:3."
"Children's book illustration of a fox and a rabbit sharing a forest picnic — flat vector style, soft pastel palette, square."
"Isometric cyberpunk city map in hand-drawn line art style — neon color accents on dark ground, square format."
"Mid-century modern travel poster for Kyoto in spring — cherry blossoms, pagoda silhouette, two-color risograph print effect, 2:3."
"Character sheet for a cozy fantasy baker — front, side, and 3/4 view, soft ink lines, muted earth tone palette, square."
"Place this perfume bottle on wet marble with soft morning light — minimalist studio, white background, 3:4 for product detail page."
"Re-shoot the sneaker on a concrete rooftop at golden hour — lifestyle scene, crop to 1:1 for Instagram grid."
"Generate 6 background variants for this face cream: studio white, moody dark, pink botanical, spa tile, linen texture, outdoor moss."
"Create a Christmas scene for this candle product — pine cones, warm candlelight, snow on window, gift ribbon, square."
"Take this jewelry photo and produce 4 lifestyle scene options: editorial magazine, luxury hotel vanity, outdoor garden, minimalist desk."
"Re-shoot every image in /products with a clean white background and consistent soft studio lighting — save all to Drive."
"Full-bleed campaign visual for a fitness brand — athlete mid-sprint, stadium lighting, bold typography space on right, 16:9."
"Split-screen ad: dull living room on the left, same room transformed with warm afternoon light and minimal furniture on the right — square."
"3 ad header variants for a back-to-school sale — bright Gen Z aesthetic, mix of flat and photographic, 1200x628px."
"Premium automotive ad visual — sports car on a winding mountain road at blue hour, drone-style wide shot, 16:9."
"Valentine's Day campaign hero — luxury chocolate box, scattered rose petals, gold accents, premium editorial feel, 16:9."
"Generate 5 display banner sizes for this campaign visual — 300x250, 728x90, 160x600, 320x50, 1200x628 — maintain composition."
"Animate this product bottle into a 10-second video — slow 360-degree rotation on white, soft studio lighting throughout."
"Create a 5-second cinemagraph from this coffee cup photo — steam rising gently, everything else perfectly still."
"15-second brand film: product dissolving into its raw ingredients, then reassembling — clean white studio, macro close-ups."
"Animate this lifestyle shot into a 7-second Stories video — slow zoom in and subtle lighting shift, 9:16."
"10-second unboxing sequence: package opens to reveal product, confetti falls — vibrant, social-native energy."
"Take hero.mp4 and extend it by 3 more 7-second segments — maintain the visual style and motion continuity."
"5 Story backgrounds for a wellness brand: soft gradient, botanical overlay, pastel textures — 9:16, top third clear for text."
"TikTok thumbnail for a cooking video — close-up of finished dish, warm tones, bold shadow space at bottom for title, 9:16."
"LinkedIn banner for a B2B SaaS company — professional team in open office, city skyline through floor-to-ceiling windows, 4:1."
"Reels cover for a travel account — golden hour beach shot with silhouette figure walking away, cinematic crop, 9:16."
"4 Pinterest pin images for a home decor brand: bedroom flat lay, kitchen counter, living room corner, bathroom shelf — 2:3 each."
"Generate this week's 7 Instagram post visuals from brief.md — maintain brand palette, vary composition, square format."
Paste a command — or ask your agent to do it for you. Sign in. Ask for what you want. That's the whole setup.
Paste one command into any terminal — or ask your AI agent (Claude Code, Codex, Cursor, Windsurf) to do it for you. Works on macOS, Linux, and Windows.
Install @picsart/gen-ai-mcp, then add gen-ai-mcp as the MCP command in your agent host.
Paste one command into any terminal — or ask your AI agent (Claude Code, Codex, Cursor, Windsurf) to do it for you. Works on macOS, Linux, and Windows.
Run gen-ai login. A browser tab opens, you click "allow", you're authenticated. Credentials stay on your machine.
Ask your agent to run gen-ai_whoami. If auth is missing, pass PICSART_TOKEN through the host config or sign in with the CLI once.
Run gen-ai login. A browser tab opens, you click "allow", you're authenticated. Credentials stay on your machine.
Run gen-ai generate and describe what you want in plain English. Your file lands in your folder, ready to use.
Describe what you want in plain English — image, video, or audio. Your agent calls the right Picsart model and saves the output for you.
Run gen-ai generate and describe what you want in plain English. Your file lands in your folder, ready to use.
Each skill is a drop-in folder that teaches your AI agent a complete job — generate a hero, cut a reel, re-shoot a 1000-SKU catalog. Install one, ask in plain English, and the agent delivers end-to-end. Below: 6 starter skills covering every media type, plus 6 persona-specific workflow packs.
Run one command in your terminal. Claude Code pulls the skill directly into your workspace — no unzipping, no manual file placement.
claude plugin marketplace add PicsArt/gen-ai-skills
Or download a specific skill as a ZIP and drop it into Claude Code — say "Install this skill for me" and it handles the rest.
The first time a skill runs, a browser tab opens and you click Allow. Credentials stay local. You never see gen-ai login — the skill handles it.
The skill auto-triggers when your prompt matches. "Generate a campaign hero, 16:9, warm tones" — the agent picks the model and runs it.
Click any skill card below. A .zip saves to your Downloads folder — instructions and any helper assets bundled together.
Unzip and move the skill folder to .cursor/skills/ (Cursor) or your project root /skills/ (Windsurf). Reload the workspace.
The skill auto-triggers on matching prompts — "render a product hero, white background, square" — no commands, no config.
Run one command. The skills package pulls all Picsart skills into your agent's skills directory automatically.
npx skills add PicsArt/gen-ai-skills
The skill opens a browser tab for OAuth on first trigger. Credentials stay local. No gen-ai login required from your side.
Skills auto-trigger on matching prompts. Works with any agent that supports the skills spec: Codex, Hermes, OpenClaw, and more.
gen-ai login && gen-ai generate -m flux-2-pro -p "sunrise over mountains"
gen-ai generate -m flux-2-pro -p "editorial hero image" --ar 5:3
gen-ai generate -m kling-v3-pro --start-frame hero.webp --duration 10 --generate-audio
gen-ai generate -m eleven-v3 -p "warm, confident voiceover"
gen-ai batch run products.json -c 4 --save-to-drive
gen-ai image -m flux-2-pro -p "product launch hero" --ar 5:3
gen-ai generate -m kling-v3-pro --start-frame portrait.png --ar 9:16 --duration 8 --generate-audio
gen-ai batch run campaign.json -c 4
gen-ai generate -m recraftv4 -i hero.webp -n 50 --ar 1:1
gen-ai batch run localize.json -c 4
gen-ai generate -m flux-2-pro -p "$HERO" --ar 5:3
gen-ai generate -m flux-2-pro -p "fitness coach, gen-z, neon vibe — full-body 4-angle casting card" --ar 1:1
gen-ai generate -m flux-2-pro -p "$PRODUCT — magazine editorial" --ar 5:3
gen-ai generate -m ideogram-character -i art.png -p "tshirt, mug, poster, hoodie, tote, sticker" -n 6
gen-ai generate -m ideogram-character -i selfie-hd.png -p "linkedin, id, editorial, casual" -n 4
pbpaste | gen-ai generate -m recraftv4 -s -o inline.webp
gen-ai generate -m kling-v3-pro --start-frame portrait.png --ar 9:16 --duration 8 --generate-audio
gen-ai generate -m recraftv4 -p "$TITLE — brand style" --ar 16:9 --script | jq -r .url
gen-ai generate -m recraftv4_vector -i style.svg -p "icon-search, icon-bell, empty-inbox" -n 6
gen-ai generate -m recraftv3-replace-bg -i raw.png -p "soft studio gradient, subtle shadow"
gen-ai generate -m flux-2-pro -p "premium website hero for $PRODUCT, clean layout, strong focal visual" --ar 16:9
gen-ai generate -m recraftv4 -p "avatar, $USER_SEED" --ar 1:1 --script | jq -r .url
gen-ai batch run catalog.json -c 10 --save-to-drive
gen-ai generate -m flux-2-pro -p "$PROMPT" --ar 5:3
gen-ai batch run press.json --save-to-drive
gen-ai config set registry.flux [email protected]
gen-ai generate -m recraftv3-replace-bg -i product.png -p "airy studio, warm natural light"
gen-ai generate -m qwen-image-edit-plus -i hero.png -p "red, navy, olive, cream" -n 4
gen-ai generate -m runway-gen4-ref -i product.png -p "on a marble kitchen counter, morning light"
gen-ai batch run holiday.json -c 4
gen-ai generate -m recraftv4 -p "$CLIENT — 5 visual directions" -n 5 --ar 16:9
gen-ai batch run pitch.json -c 4
gen-ai batch run clients.json -c 4
gen-ai download --folder "client-x" --out ./client-x-deliverable
gen-ai generate -m kling-v3-pro --start-frame talent.png --ar 9:16 --duration 8 --generate-audio
Every visual below was generated with the CLI or the MCP — no mock-ups. Click any card to reveal the exact prompt, or copy it and run it yourself.
Every visual below was generated by an AI agent calling Picsart MCP — no mock-ups. Click any card to reveal the prompt, or copy it and ask your own agent.
Every visual below was generated with the CLI or the MCP — no mock-ups. Click any card to reveal the exact prompt, or copy it and run it yourself.
$ gen-ai generate -m recraftv4 -p "editorial still-life..." --ar 1:1
$ gen-ai generate -m kling-v3-standard -p "neon liquid..." --duration 5
$ gen-ai generate -m eleven-v3 -p "Welcome to Picsart..."
$ gen-ai generate -m recraftv4 -p "campaign asset kit..."
$ gen-ai generate -m recraftv4 -p "portrait silhouette..."
$ gen-ai generate -m recraftv4 -p "URL to OG preview..."
$ gen-ai generate -m recraftv4 -p "branded product assets..."
$ gen-ai generate -m recraftv4 -p "workflow visualization..." --ar 16:9
Built on the same stack that powers Picsart Agents — now open to every builder.
gen-ai generate.brand.md. Off-brand output auto-rejected.The @picsart/gen-ai-mcp server plugs into Claude Code, ChatGPT, Cursor, Windsurf, Zapier, n8n — any MCP-compatible host. Your agent gets native access to all 150+ Picsart models and can call them in plain English.
On macOS or Linux: curl -fsSL https://picsart.com/gen-ai-cli/install.sh | bash. On Windows PowerShell: irm https://picsart.com/gen-ai-cli/install.ps1 | iex. Or via npm: npm install -g @picsart/gen-ai. Requires Node.js 22+.
150+ frontier models across 25+ providers — Flux 2 Pro/Max, Sora 2, Kling 3.0, Veo 3.1, Runway Gen 4.5, Luma Ray 2, Nano Banana, Imagen 4, Recraft V4, Seedream 5, Ideogram v3, Hunyuan V3, Hailuo 2.3, Pika, ElevenLabs, MiniMax Music, and more. Covers image, video, and audio generation.
Run gen-ai models list to see everything available on your account.
Yes — any agent with shell access can already call gen-ai directly. Claude Code, Codex, and Cursor happily run CLI commands today; drop one of our skill files into your agent and it'll start generating media on request.
A dedicated @picsart/gen-ai-mcp server is available now, exposing every CLI command as a native MCP tool call for Claude Code, Cursor, VS Code Copilot, Windsurf, ChatGPT, and any MCP host.
Claude Code uses claude mcp add gen-ai-mcp -- gen-ai-mcp. Codex uses codex mcp add gen-ai-mcp -- gen-ai-mcp. Cursor, Windsurf, and VS Code can use gen-ai-mcp as the MCP command in their config.
Claude Code, Cursor, Codex, Windsurf, VS Code Copilot, ChatGPT with compatible connector setup, and any host that implements the Model Context Protocol can run the Picsart MCP server.
No. The MCP server is standalone and runs as its own gen-ai-mcp process. The CLI is optional for terminal, CI, and batch workflows. Both use the same Picsart account and credit balance.
In Claude Code: claude plugin marketplace add PicsArt/gen-ai-skills. Via npx: npx skills add PicsArt/gen-ai-skills. Or download the ZIP from picsart.com/gen-ai-skills/ and place it in ~/.claude/skills/. Authentication runs automatically on first use — no separate login needed.
150+ frontier models across 25+ providers — Flux 2 Pro/Max, Sora 2, Kling 3.0, Veo 3.1, Runway Gen 4.5, Luma Ray 2, Nano Banana, Imagen 4, Recraft V4, Seedream 5, Ideogram v3, Hunyuan V3, Hailuo 2.3, Pika, ElevenLabs, MiniMax Music, and more. Covers image, video, and audio generation.
Run gen-ai models list to see everything available on your account.
Yes — any agent with shell access can already call gen-ai directly. Claude Code, Codex, and Cursor happily run CLI commands today; drop one of our skill files into your agent and it'll start generating media on request.
The @picsart/gen-ai-mcp server is available now. Add it alongside Skills to get native MCP tool calls in Claude Code, Cursor, Windsurf, VS Code Copilot, and ChatGPT.
The CLI is free. Generations consume Picsart credits, drawn from a single balance that spans every model and modality. New accounts get a starter balance; run gen-ai pricing -m <model> before heavy usage to confirm cost. One key, one balance, one invoice — no per-vendor contracts. The MCP server shares the same credit balance.
Yes. Point --input-dir at a folder of images, or run a manifest JSON with per-item models and prompts. Use --concurrency to tune throughput, --resume to pick up where a run left off, and --script to emit JSON on stdout for CI/CD pipelines.
Pass --rules brand.md to gate every generation through your brand policy — colors, typography, tone, banned terms, mandatory disclaimers. Off-brand output is auto-rejected and logged to an audit trail so you can trace every asset back to its prompt, model, and rule set.
The gen-ai CLI is the same stack that powers Picsart Agents. Everything Picsart Agents does — catalog re-shoots, brand-gated generation, multi-platform campaign kits — is now open to any builder with a terminal. The @picsart/gen-ai-mcp server is live, so the same stack plugs into any MCP host too.
macOS 12+ (Apple Silicon and Intel), Linux (x86_64 and ARM64), and Windows 10/11. macOS and Linux install a signed, self-contained binary (~60 MB) via curl — no Node.js required. Windows and the npm path install the pure-JS distribution and need Node.js 22+ (the PowerShell installer sets it up automatically via winget). Runs in Docker, GitHub Actions, GitLab CI, and WSL.
Run gen-ai login once — the CLI opens a browser for OAuth and stores a token in your OS keychain. For CI, set PICSART_API_KEY as an environment variable. When the MCP server ships, hosts will inherit the same credentials with no per-agent setup.
Outputs are written to ./out/ by default, or wherever you point --output. Nothing is uploaded to Picsart beyond the prompt and input assets needed to run the model. You own the outputs; enterprise plans add retention controls, SSO, and data-residency options.
Yes — that's the whole point. Every command takes -m <model>, flags are consistent across providers, and stdin/stdout piping works the same way everywhere. Swap Flux for Recraft, or Sora for Veo, by changing one flag. No per-vendor SDKs, no rewriting prompts.
Learn how to build reusable creative AI skills step by step.
Claude 
Claude Code 
Cursor 
Windsurf 
Zapier 
n8n 
Make 
Gemini 
