# Lightdrift [Verified owner]

**Category:** 🔎 Search & Data Extraction  
**Repository:** https://lightdrift.ai/mcp  
**Views:** 10  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/lightdrift

## Description
Semantic image search for presentation, website and content workflows. Search images, find similar images, and retrieve files with provenance and license information. OAuth and credit required for paid searches. Review source rights and attribution. Try the presentation example and reproducible terminal walkthrough: https://github.com/JacksonHolland/lightdrift-claude-plugin

## Claude Desktop Quick Installation
Remote MCP endpoint (confidence: medium). Install path inferred — verify against the README before running it. Add as a URL/SSE server in your client:

```json
"mcpServers": {
  "lightdrift": {
    "url": "https://lightdrift.ai/mcp"
  }
}
```

## Documentation

## What the Lightdrift MCP server does

Lightdrift MCP server supports image discovery for presentations, websites, and other content workflows. It lets an agent search for images by meaning, locate images similar to a reference, and retrieve image files. Retrieved results include provenance and licensing information so developers can review where an asset came from and what attribution or usage considerations apply.

The service belongs in the 🔎 Search & Data Extraction category when an application needs more than a keyword-only image lookup. It is relevant for workflows that select visual assets during content creation, especially when source information must remain attached to the result.

## How it works

An MCP-compatible agent sends image-search or similarity requests to Lightdrift. Search requests return candidate images based on semantic intent, while similarity requests help identify images related to an existing image. File retrieval provides the asset together with provenance and license details when those are available through the service.

Lightdrift MCP server can be used in presentation, website, and general content workflows. Paid searches require OAuth authentication and available credits.

## Setup and configuration

The provided source identifies Lightdrift at https://lightdrift.ai/mcp, but it does not include installation commands, configuration fields, environment variables, or transport details. Set up therefore requires consulting the linked service documentation rather than relying on assumptions about a local package or command.

A presentation example and a reproducible terminal walkthrough are available through the linked GitHub project at https://github.com/JacksonHolland/lightdrift-claude-plugin.

## Tools and capabilities

Lightdrift MCP server is described as providing these capabilities:

- Search images using semantic intent.
- Find images similar to a supplied image or search result.
- Retrieve image files for downstream content workflows.
- Return provenance information with retrieved files.
- Return license information for rights and attribution review.

## Limitations and notes

Review source rights and attribution before using retrieved images. Provenance and license information supports that review but does not replace a developer's responsibility to confirm whether an asset can be used for the intended purpose.

OAuth and credits are required for paid searches. The available description does not identify which searches are paid, how credits are consumed, or whether a free tier exists. It also does not document rate limits, file formats, result ranking, supported image sources, or client-specific behavior.

_Full upstream README: https://allmcps.com/mcp/lightdrift/readme_