MCP Servers/Tastepix
T

Tastepix

Official

Hosted remote MCP server for image generation and editing: 12 models (Nano Banana, GPT Image, FLUX, Seedream, Recraft and more) inside Claude, ChatGPT, Cursor, Claude Code and Codex. No API keys, pay per image.

Tastepix is a hosted remote MCP server that adds image generation and editing to the AI assistants developers already use. Once connected, you ask for an image in plain language and it appears inline in the chat. The assistant writes the prompt and picks the model, or you name one yourself, such as Nano Banana Pro, GPT Image, FLUX, Seedream, Recraft or Qwen Image.

Beyond text-to-image, the server exposes tools for editing uploaded images, combining references, batches of up to four images, background removal, upscaling and image-to-SVG vectorizing. Every result is also saved to a private gallery at tastepix.com.

There are no provider API keys to manage. You sign in once with OAuth, keep a prepaid balance and pay per image, with prices shown per model on the Tastepix site. Failed generations are refunded automatically, and new accounts get a small free credit. The server is listed in the Official MCP Registry as io.github.paablobello/tastepix.

What You Can Do

Generate hero images, icons and illustrations for a project without leaving Claude Code or Cursor.
Compare how different image models handle the same prompt by naming a model in chat.
Edit an uploaded photo, remove its background or upscale it from inside ChatGPT or Claude.
Vectorize a generated logo or icon into SVG for use in a web app.
Create small batches of image variations for A/B testing marketing pages.

Setup Guide

1
Server URL: https://mcp.tastepix.com/mcp (Streamable HTTP, OAuth sign-in)
2
Claude Code: claude mcp add --transport http --scope user tastepix https://mcp.tastepix.com/mcp, then run /mcp and sign in
3
Cursor: add { "mcpServers": { "tastepix": { "url": "https://mcp.tastepix.com/mcp" } } } to ~/.cursor/mcp.json
4
Codex: codex mcp add tastepix --url https://mcp.tastepix.com/mcp, then codex mcp login tastepix
5
Claude or ChatGPT: add a custom connector with the server URL and sign in to Tastepix
6
Ask your assistant for an image to test the connection

How to Connect in Antigravity

To connect to a custom MCP server in Antigravity:

1
Open the MCP store via the "..." dropdown at the top of the editor's agent panel.
2
Click on "Manage MCP Servers"
3
Click on "View raw config"
4
Modify the mcp_config.json with your custom MCP server configuration.

Here is an example configuration compatible with MCP-enabled editors like Antigravity, Windsurf, Claude Desktop, and Cursor:

{
  "mcpServers": {
    "supabase-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "@supabase/mcp-server-supabase@latest",
        "--access-token",
        "add-token-here"
      ],
      "env": {}
    },
    "github-mcp-server": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "GITHUB_PERSONAL_ACCESS_TOKEN",
        "ghcr.io/github/github-mcp-server"
      ],
      "env": {
        "GITHUB_PERSONAL_ACCESS_TOKEN": "add-your-token-here"
      }
    }
  }
}

This configuration defines two servers: one running via npx (Supabase) and another via docker (GitHub). Make sure to replace the placeholder tokens with your actual API keys.

For a detailed step-by-step walkthrough, check out our MCP Tutorial or explore more in our Top 10 Essential MCP Servers guide.

How to Connect in Cursor

To add this MCP server to Cursor:

1
Open Cursor Settings (File > Preferences > Cursor Settings).
2
Navigate to Features > MCP Servers.
3
Click on "Add new MCP server".
4
Enter the Name, Command, and Arguments as specified in the configuration above.

For more details, refer to the official Cursor MCP documentation.

Sponsors
Agent SkillsThe skill registry for AI agentsSponsoredRunableBuild, run and grow — one AI agent does it allSponsored
Advertise here
4/6 spots left

Maintained By

Tastepix

Tastepix, created by Pablo Bello, is a pay-per-image service that brings multiple image models into chat assistants through a single hosted MCP server.

Visit Website

Related Servers