coreviz

@coreviz/cli

An AI-powered CLI for working with photos. Semantically search, edit, tag and generate metadata for thousands of photos right from the command line.

Installation

npm install -g @coreviz/cli

Usage

# Run directly with npx
npx @coreviz/cli [command]

# Or if installed globally
coreviz [command]

Commands

AI Features

Describe an image:

npx @coreviz/cli describe path/to/image.jpg

Edit an image with a text prompt (🍌 Nano Banan + Flux Kontext in the CLI!):

npx @coreviz/cli edit path/to/image.jpg --prompt "make it cyberpunk style"

Generate tags or classify an image:

npx @coreviz/cli tag path/to/image.jpg "objects in the image"

Classify an image using specific choices:

npx @coreviz/cli tag path/to/image.jpg --choices "receipt,invoice,document" --single

Run tagging locally (offline capable):

npx @coreviz/cli tag path/to/image.jpg "prompt" --mode local

Search local images using natural language:

npx @coreviz/cli search "a person wearing a red t-shirt"

This will index the images in your current directory (creating a .index.db file) and return the top matches for your query.

You can also use the cloud API for embeddings:

npx @coreviz/cli search "query" --mode api

Scripting

All commands support a --quiet flag to suppress UI output and return only raw results.

# Returns only the file path of the edited image
npx @coreviz/cli edit image.jpg "prompt" --quiet

Authentication

Login to CoreViz using device authorization:

npx @coreviz/cli login

Logout:

npx @coreviz/cli logout

Check login status:

npx @coreviz/cli whoami

MCP Server (Claude Code Integration)

@coreviz/cli includes a built-in MCP server that exposes your CoreViz visual library as tools for Claude Code and other MCP-compatible AI agents — turning CoreViz into a visual memory for your AI workflows.

Setup

1. Login (if you haven't already):

   npx @coreviz/cli login
   

2. Connect to your MCP client:

Claude Code — Install the plugin (recommended):

claude plugin marketplace add coreviz/cli
claude plugin install coreviz@coreviz

Or configure MCP manually in ~/.claude/settings.json:

{
  "mcpServers": {
    "coreviz": {
      "command": "npx",
      "args": ["coreviz-mcp"]
    }
  }
}

Claude Desktop — Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "coreviz": {
      "command": "npx",
      "args": ["coreviz-mcp"]
    }
  }
}

3. In Claude Code, run /mcp to confirm the server is connected.

Available MCP Tools