Picsart Creative APIs MCP

Our MCP server exposes the same functionality as our APIs, but through the Model Context Protocol. This allows AI assistants to directly interact with your API. On top of that, one can use this same server to get help and guidance with the documentation, e.g. get integration code written, etc.

MCP Endpoint

wss://mcp.picsart.io/v1 or https://mcp.picsart.io/v1

Authentication

• Requires a Picsart API key (same key you use for REST APIs).
• Send the key via the x-picsart-api-key request header.
• Token can be generated from your dashboard.

What's included

• 90+ dedicated tools — one per API action (endpoint)
• 5 generic tools for documentation and coding
• Each API action (remove background, upscale, text-to-image, etc.) is exposed as its own tool. You call one tool and get the result — no discovery step needed.

How to Connect

{
  "mcpServers": {
    "picsart": {
      "type": "http",
      "url": "https://mcp.picsart.io/v1",
      "headers": {
        "X-Picsart-API-Key": "<your_api_key>"
      }
    }
  }
}

{
  "mcpServers": {
    "picsart": {
      "type": "http",
      "url": "https://mcp.picsart.io/v1",
      "headers": {
        "X-Picsart-API-Key": "<your_api_key>"
      }
    }
  }
}

{
  "mcpServers": {
    "picsart": {
      "type": "http",
      "url": "https://mcp.picsart.io/v1",
      "headers": {
        "X-Picsart-API-Key": "<your_api_key>"
      }
    }
  }
}

{
  "inputs": [
    {
      "type": "promptString",
      "id": "picsart-api-key",
      "description": "Picsart API Key",
      "password": true
    }
  ],
  "servers": {
    "picsart": {
      "type": "http",
      "url": "https://mcp.picsart.io/v1",
      "headers": {
        "x-picsart-api-key": "${input:picsart-api-key}"
      }
    }
  }
}

{
  "mcpServers": {
    "picsart": {
      "type": "http",
      "url": "https://mcp.picsart.io/v1",
      "headers": {
        "x-picsart-api-key": "<YOUR_API_KEY>"
      },
      "tools": ["*"]
    }
  }
}

{
  "mcp": {
    "servers": {
      "picsart": {
        "url": "https://mcp.picsart.io/v1",
        "transport": "streamable-http",
        "headers": {
          "X-Picsart-API-Key": "<your_api_key>"
        }
      }
    } 
  } 
}

PICSART_API_KEY=<your_api_key>

mcp_servers:
  picsart:
   url: "https://mcp.picsart.io/v1"
   headers:
     X-Picsart-API-Key: ${PICSART_API_KEY}

hermes chat

Testing Your MCP Setup

Once configured, you can test your MCP server connection:

1. Open your AI editor (Cursor, Windsurf, etc.)
   1. VS Code: Open Copilot Chat → select Agent → open the tools list (tools icon) and ensure Picsart is enabled before running prompts.
   2. OpenClaw and Hermes Agent users can run the same test prompts in their normal chat flow after configuring the MCP server.
2. Start a new chat with the AI assistant
   1. Copilot CLI: Run /mcp show picsart to confirm the server is running and tools are available, then run the prompts below.
3. Ask about Picsart Creative APIs - try questions like:
   • "Use picsart. Remove background from https://pastatic.picsart.com/31399405932227146498.png"
   • "Use picsart. Get width and height of https://dcdn.picsart.com/dev/a61e8550-d4c5-4837-a10d-cc6a0b3cd971.mov"
   • "Use picsart. How do I remove backgrounds from my product images?"
   • "Use picsart. Show me an example of how images can be upscaled"
   • "Use picsart. Check my credits balance (image-credits-balance) and confirm auth is working."
   • "Use picsart. Remove background from <image_url>, then generate a TypeScript function + a Jest test that mocks the API call."
   • "Use picsart. Upscale <image_url>, then generate a Python example + a pytest test with error-handling."

Available Tools

The available tools is a reflection of how API services are grouped under the API Reference:

image-remove-background (Remove & Change Background), image-upscale (Upscale), image-ultra-upscale (Ultra Upscale), image-ultra-upscale-getresult (Get the Ultra Upscale Result), image-ultra-enhance (Ultra Enhance), image-face-enhance (Face Enhancement), image-list-effect-names (Effect Names), image-apply-effect (Effects), image-create-effect-previews (Effect Previews), image-apply-laser-engraving-effect (Laser Engraving Effect), image-list-ai-effect-names (AI Effect Names), image-apply-ai-effect (AI Effects), image-adjust (Adjust), image-transfer-color (Color Transfer), image-transfer-style (Style Transfer), image-create-mask-previews (Mask Previews), image-apply-mask (Masks), image-generate-texture (Texture Generator), image-generate-pattern (Pattern Generator), image-vectorize-raster-to-svg (Image Vectorizer), image-design-import (Design Import), image-design-import-getresult (Get the Design Import Result), image-surfacemap (Surfacemap Image), image-add-watermark (Add Watermark), image-upload (Upload Image), image-credits-balance (Credits Balance), image-edit (Basic Editing), image-blend (Blending), image-tagging (Image Tagging), image-describer (Describe Image), image-cars-classifier (Classify the Car Image)

video-upscale-fps (Video FPS Upscale), video-upscale-fps-getresult (Get the Video FPS Upscale Result), video-adjust (Adjust Video), video-apply-effect (Video Effects), video-edit (Edit Video), video-trim (Trim Video), video-crop (Crop Video), video-concat (Concatenate Videos), video-concat-segments (Create a Video Highlight), video-fit (Fit Video), video-metadata (Get Video Metadata), video-get-thumbnail (Get Video Thumbnail), video-set-thumbnail (Set Video Thumbnail), video-set-thumbnail-getresult (Get the Video with Updated Thumbnail), video-encode-ctv (Make Video CTV-Compatible), video-adjust-audio (Adjust Audio), video-transcribe-audio (Transcribe Audio), video-transcribe-audio-getresult (Get the Transcribe Audio Result), video-add-watermark (Add Watermark), video-getresult (Get the Video Result), video-upload (Upload Files), video-credits-balance (Credits Balance)

genai-text2text-completions (Text2Text Completions), genai-text2image (Text2Image), genai-text2image-getresult (Get the Text2Image Result), genai-text2sticker (Text2Sticker), genai-text2sticker-laser-engraving (Text2Sticker with Laser Engraving Effect), genai-text2sticker-getresult (Get the Text2Sticker Result), genai-generate-logo (Logo Generator), genai-generate-logo-getresult (Get the Logo Generator Result), genai-image-inpainting (Inpaint Image), genai-image-outpainting (Outpaint Image), genai-smart-background (Smart Background), genai-expand-image (Expand Image), genai-generate-image-bleed (Generate Bleed), genai-image-inpainting-getresult (Get the Inpainting Result), genai-credits-balance (Credits Balance)

vd-export-replay (Export Replay), vd-export-replay-getresult (Get the Export Replay Result), vd-describe-variable-data-content (Describe Replay's Variable Data), vd-export-variable-data-content (Export Variable Data Content), vd-export-variable-data-content-getresult (Get the Export Variable Data Content Result), vd-credits-balance (Credits Balance)

list-endpoints (List Endpoints), get-endpoint (Get Endpoint), get-request-body (Get Request Body), get-response-schema (Get Response Schema), list-security-schemes (List Security Schemes), search-specs (Search Specs), list-specs (List Specs), execute-request (Execute Request), get-code-snippet (Get Code Snippet)

As a rule of thumb, you can use this pattern to decide which tools to include. For instance, the name image includes all tools under Programmable Image APIs product. At the same time, if one uses image-remove-background that will include only services under the Remove Background category of Programmable Image APIs.

Usage Optimizations

There are certain limitations when running Cursor (e.g. 40-tool limit and up to 80 with settings) and similar tools. For instance, during the process you may get messages such as:

• Exceeding total tools limit. You have 81 tools from enabled servers.

In order to make sure the number of tools and tokens are used efficiently, you can do this type of optimizations:

Define which tools to include

mcp_servers:
  picsart:
    url: "https://mcp.picsart.io/v1"
    headers:
      X-Picsart-API-Key: ${PICSART_API_KEY}
    tools:
      include:
        - image-remove-background
        - image-upscale
        - genai-text2image