Agentic Image Optimization in Windsurf

Add tinify.ai to Windsurf's MCP configuration and optimize images from Cascade AI. Uses the MCP Server surface.

Requirements

• Node.js ≥ 18
• Windsurf IDE

Install

Open Windsurf Settings → MCP Servers → Add Server, or edit ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "tinify-ai": {
      "command": "npx",
      "args": ["-y", "@tinify-ai/mcp-server"]
    }
  }
}

Restart Windsurf after saving.

Verify

Open Windsurf Settings → MCP Servers. The tinify-ai server should show as Running.

Usage example

In Cascade AI:

Optimize all images in /public/assets — compress and convert to WebP

Cascade calls optimize_image for each file and reports savings.

Auth

Guest mode: 20 credits/day, no signup. To use account credits:

Log in to tinify.ai

Pricing →

Troubleshooting

Server shows "Stopped" in MCP settings

Check that Node.js ≥ 18 is installed:

node --version

Config file location

Windsurf MCP config is at ~/.codeium/windsurf/mcp_config.json. Project-level configs are not supported — use the global config.

Insufficient credits error

20 free credits/day, resets at midnight UTC. Upgrade →

← MCP Server overview