HomeNews

Claude AI Now Controls Photoshop via MCP Integration

Naman
Naman
•
Oct 24, 2026 3:50 AM
•
0
 min read
Select Emergent as your Preferred news source
Claude AI Now Controls Photoshop via MCP Integration

💡 TL;DR

  • Developer kathan3009 released Photoshop MCP, an open-source server that gives Claude AI direct control over Adobe Photoshop editing functions.
  • The integration uses Anthropic's Model Context Protocol to translate natural language commands into executable Photoshop actions and layer manipulations.
  • Now available on GitHub, the tool enables designers to automate complex workflows by describing edits conversationally instead of manual clicking.

A developer has released an open-source integration that gives Claude AI direct control over Adobe Photoshop, enabling designers to manipulate images through conversational commands. The Photoshop MCP server, officially launched on October 23, 2026, uses Anthropic's Model Context Protocol to bridge natural language instructions with Photoshop's automation capabilities.

The project emerged on Hacker News as a Show HN demonstration, showcasing how large language models can interface with professional creative software through standardized protocols. Unlike basic plugin architectures, the MCP implementation provides Claude with programmatic access to Photoshop's core functions, from layer management to filter application.

How the Integration Works

The Photoshop MCP server acts as a middleware layer between Claude and Adobe's scripting engine. When a user asks Claude to perform an editing task, the AI translates the request into ExtendScript commands that Photoshop can execute natively. This architecture preserves the full power of Photoshop's automation API while abstracting away the complexity of scripting syntax.

Key capabilities include layer creation and manipulation, selection tools, color adjustments, and filter application. The system maintains context awareness across multi-step workflows, allowing users to refine edits through follow-up prompts without starting over. Error handling built into the MCP server provides Claude with feedback when operations fail, enabling the AI to suggest alternative approaches.

Technical Implementation Details

The server runs locally on the designer's machine, establishing a communication channel between Claude's API and Photoshop's scripting interface. Authentication and file access permissions remain under user control, with Claude unable to access files or perform actions outside explicitly granted scope.

  • Supports Photoshop CC 2020 and later versions with ExtendScript compatibility
  • Implements rate limiting to prevent resource exhaustion during batch operations
  • Includes example prompts for common design tasks like background removal and color grading
  • Provides logging functionality for debugging failed commands

The open-source repository includes setup instructions for both macOS and Windows environments, along with a configuration file that defines which Photoshop features Claude can access. Developers can extend the base functionality by adding custom ExtendScript modules to the MCP server's command registry.

Availability and Access

The Photoshop MCP server is officially released on October 23, 2026, and available immediately through GitHub. The project ships under an MIT license, allowing commercial use and modification. Installation requires Node.js 18 or higher, an active Claude API subscription, and a licensed copy of Adobe Photoshop.

Developer kathan3009 maintains the repository as an independent project, not affiliated with Anthropic or Adobe. The initial release focuses on core editing primitives, with the roadmap indicating plans for generative fill integration and batch processing workflows in future versions.

What This Means

This integration represents a practical application of the Model Context Protocol beyond chatbot interfaces, demonstrating how AI assistants can extend into specialized professional tools. For designers, the ability to describe edits conversationally rather than memorizing menu paths could accelerate iterative workflows, particularly for repetitive tasks like asset resizing or color correction across multiple files. The open-source nature of the project also creates opportunities for the community to build domain-specific editing vocabularies tailored to industries like e-commerce product photography or social media content creation.

About the writer

A growth marketer with varied interests and the proven ability to acquire new skills fast. Currently engrossed in all things AI.

Start Building
on Emergent today
Try Emergent