# Picsart gen-ai CLI, MCP & Skills > Developer docs for driving Picsart's AI catalog — 262 models from 32 providers across image (79), video (103), audio (32), text (45), json (3) — from your terminal with the gen-ai CLI, or from any AI agent via Skills and the hosted MCP server. CLI and MCP model availability depends on the installed or deployed version. Models have stable IDs, but execution methods and availability differ by interface. Use `gen-ai generate -m ` or `picsart_generate` for supported media models. Structured JSON models use the SDK’s `ai.run()` method; verify CLI, MCP, and Playground support separately. The MCP server is hosted at https://api.picsart.com/gen-ai/mcp: agents connect over HTTP and sign in with Picsart OAuth in the browser, so it needs no CLI install, no `gen-ai login` and no API key. Skills run the gen-ai CLI and need `gen-ai login`. For a point-and-click web UI instead, see the AI Playground app (linked under Optional). ## Getting started - [Introduction](https://picsart.github.io/picsart-mcp-cli-docs/guide/introduction): what this is and the three interfaces - [Installation](https://picsart.github.io/picsart-mcp-cli-docs/guide/installation): install the gen-ai CLI, add Skills, or connect the hosted MCP server - [SDK 6.29.0 catalog update](https://picsart.github.io/picsart-mcp-cli-docs/guide/sdk-6-29-update): 3 additions, 4 removals, and structured JSON output - [Authentication](https://picsart.github.io/picsart-mcp-cli-docs/guide/authentication): `gen-ai login` for the CLI and Skills; in-agent OAuth for MCP; API key for the SDK ## Interfaces - [CLI Quickstart](https://picsart.github.io/picsart-mcp-cli-docs/guide/cli-quickstart): generate, batch, pipe, upload from the terminal - [MCP Quickstart](https://picsart.github.io/picsart-mcp-cli-docs/guide/mcp-quickstart): connect the hosted MCP server (https://api.picsart.com/gen-ai/mcp) to any MCP-compatible agent; nothing to install - [Skills](https://picsart.github.io/picsart-mcp-cli-docs/guide/skills): drop-in skills for Claude Code, Cursor, Windsurf, ChatGPT ## Picsart Media Studio (separate MCP connector) - [Media Studio overview](https://picsart.github.io/picsart-mcp-cli-docs/guide/media-studio/): build video, slideshows and captions from your own footage and templates; remote connector, added and signed in separately from the gen-ai MCP server - [What you can do](https://picsart.github.io/picsart-mcp-cli-docs/guide/media-studio/tools): all 29 picsart_media_* tools and what each is for - [Troubleshooting](https://picsart.github.io/picsart-mcp-cli-docs/guide/media-studio/troubleshooting): connecting, signing in, files from your computer, blank text, credits, previews ## Concepts - [Generating media](https://picsart.github.io/picsart-mcp-cli-docs/guide/generating): prompts, inputs, aspect ratio, duration, outputs - [Files & Drive](https://picsart.github.io/picsart-mcp-cli-docs/guide/files-and-drive): upload, download, and list files on Picsart Drive - [Pricing & Credits](https://picsart.github.io/picsart-mcp-cli-docs/guide/pricing): per-generation credit cost; estimate before running - [Batch & Automation](https://picsart.github.io/picsart-mcp-cli-docs/guide/batch): manifest runner, from-dir, auto-download ## Model reference - [Model Reference overview](https://picsart.github.io/picsart-mcp-cli-docs/reference/): all 262 models, how to read a provider page - [Model Catalog](https://picsart.github.io/picsart-mcp-cli-docs/reference/catalog): searchable, filterable catalog of every model - [Image generation](https://picsart.github.io/picsart-mcp-cli-docs/reference/image): 79 image models - [Video generation](https://picsart.github.io/picsart-mcp-cli-docs/reference/video): 103 video models - [Audio generation](https://picsart.github.io/picsart-mcp-cli-docs/reference/audio): 32 audio models - [Text & analysis](https://picsart.github.io/picsart-mcp-cli-docs/reference/text): 45 text models - [Structured JSON output](https://picsart.github.io/picsart-mcp-cli-docs/reference/json): 3 JSON models ## Providers Each provider page lists its models with CLI + MCP examples, a parameter table sourced from the live catalog, and links to official vendor docs. - [All providers](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/): index of all 32 providers - [Anthropic](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/anthropic): 11 models — text - [Async](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/async): 1 model — audio - [ByteDance](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/bytedance): 24 models — image, video, audio - [Creatify](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/creatify): 2 models — video - [Depth Anything](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/depth-anything): 1 model — video - [ElevenLabs](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/elevenlabs): 16 models — audio, text - [Flux (Black Forest Labs)](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/flux): 10 models — image, video - [Google](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/google): 26 models — image, video, audio, text, json - [Grok (xAI)](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/grok): 10 models — image, video, audio, text - [Happy Horse](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/happyhorse): 5 models — video - [HeyGen](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/heygen): 3 models — video - [Hunyuan](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/hunyuan): 1 model — image - [Ideogram](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/ideogram): 8 models — image - [Kimi](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/kimi): 3 models — text - [Kling](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/kling): 16 models — image, video, audio - [LTX (Lightricks)](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/ltx): 9 models — video - [Luma](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/luma): 5 models — image, video - [Meta](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/meta): 2 models — image, video - [MiniMax](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/minimax): 21 models — image, video, audio, text - [Mirelo](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/mirelo): 3 models — video, audio - [OpenAI](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/openai): 26 models — image, text, json - [OVI](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/ovi): 1 model — video - [Picsart](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/picsart): 9 models — image, video - [PixVerse](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/pixverse): 6 models — video - [Qwen (Alibaba)](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/qwen): 2 models — image - [Recraft](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/recraft): 25 models — image - [Runway](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/runway): 4 models — image, video - [Topaz](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/topaz): 2 models — image, video - [TypeSafe](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/typesafe): 1 model — json - [VEED](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/veed): 2 models — video - [Videography](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/videography): 1 model — video - [Wan (Alibaba)](https://picsart.github.io/picsart-mcp-cli-docs/reference/providers/wan): 6 models — video ## Optional - [Changelog](https://picsart.github.io/picsart-mcp-cli-docs/changelog): recent catalog and CLI/MCP changes - [AI Playground app](https://picsart.com/ai-playground/): browser-based web UI for the same catalog