Skip to main content
Glama
spoiledbeef

Image Scraper MCP Server

by spoiledbeef
OverviewSchemaRelated ServersScoreDiscussions
Local

Image Scraper

A Python web scraper that retrieves images for a given search query. Supports DuckDuckGo Images and Google Maps place photos — no API key required, and every result carries the credit/attribution you need to safely reuse the image.

Features

  • Three scraping scripts:

    • image_scraper_selenium.py (Selenium DuckDuckGo) — recommended; returns the real image URL plus full source attribution

    • image_scraper_maps.py (Selenium Google Maps) — place photo scraper with photographer attribution

    • image_scraper.py (BeautifulSoup DuckDuckGo) — fast fallback that only returns URL + alt text

  • MCP server (image_scraper_mcp_server.py) exposing both search_images and search_maps_images tools

  • Credit & attribution built-in — every result includes the publisher domain, source page URL, article title, and image dimensions so you can credit the image instead of just receiving a proxied redirect

  • Configurable number of images, headless mode, and error handling

Related MCP server: ddg-search-mcp

Credit & Attribution

Reusing images from search results requires credit. Both scrapers are built to give you what you need:

Tool

What you get for credit

search_images (DDG)

source_domain, source_url, title, source_favicon_url, width, height — the real image URL, with DDG's iu/?u= redirector unwrapped

search_maps_images (Google Maps)

author (the Google Maps contributor's name) and author_profile_url — the photographer's profile URL

Suggested credit format:

  • DuckDuckGo results: Image: {title or alt} — {source_domain} ({source_url}). If source_domain or source_url is missing for a result, treat that entry as low-attribution and prefer a different one.

  • Google Maps results with an author: Photo by {author} ({author_profile_url}).

  • Google Maps results with no author (official place photos): Image: {Place Name} via Google Maps or Source: Google Maps.

Installation

  1. Install Python dependencies:

uv sync

  1. Make sure you have Google Chrome installed (required for Selenium). Selenium 4+ automatically manages ChromeDriver, no manual installation needed.

Usage

Selenium DuckDuckGo (recommended)

uv run python image_scraper_selenium.py "cute cats" 5

Prints a JSON payload with attribution per image:

{
  "query": "cute kittens",
  "source": "duckduckgo",
  "count": 2,
  "images": [
    {
      "url": "https://tse3.mm.bing.net/th/id/OIP.6ytt01A4fK8ToB7he6XJegHaFD?pid=Api",
      "alt": "Cute Kittens Playing With Yarn Free Stock Photo - Public Domain Pictures",
      "title": "Cute Kittens Playing With Yarn Free Stock Photo - Public Domain Pictures",
      "source_url": "https://www.publicdomainpictures.net/en/view-image.php?image=588007&picture=cute-kittens-playing-with-yarn",
      "source_domain": "publicdomainpictures.net",
      "source_favicon_url": "https://external-content.duckduckgo.com/ip3/www.publicdomainpictures.net.ico",
      "width": 1920,
      "height": 1309,
      "ddg_proxied_url": "https://external-content.duckduckgo.com/iu/?u=..."
    }
  ]
}

The url is the real image URL (DDG's external-content.duckduckgo.com/iu/?u= redirector is unwrapped). ddg_proxied_url is kept for reference. Width and height are read from the dimensions overlay on the result tile.

When called through the MCP server, the same data is rendered as human-readable text that includes a Credit: line per result, plus the source page title and dimensions:

Found 2 images for 'cute kittens':

1. https://tse3.mm.bing.net/th/id/OIP.6ytt01A4fK8ToB7he6XJegHaFD?pid=Api
   Credit: publicdomainpictures.net (https://www.publicdomainpictures.net/en/view-image.php?image=588007&picture=cute-kittens-playing-with-yarn)
   Title: Cute Kittens Playing With Yarn Free Stock Photo - Public Domain Pictures
   Dimensions: 1920 × 1309
   Alt: Cute Kittens Playing With Yarn Free Stock Photo - Public Domain Pictures

2. https://tse1.mm.bing.net/th/id/OIP.So-qlqEv3QwuZpL4cOY-PwHaHa?pid=Api
   Credit: wallpaperaccess.com (https://wallpaperaccess.com/cute-cats-and-kittens)
   Title: Cute Cats and Kittens Wallpapers - Top Free Cute Cats and Kittens ...
   Dimensions: 2560 × 1600
   Alt: Cute Cats and Kittens Wallpapers - Top Free Cute Cats and Kittens ...

Google Maps place photos

uv run python image_scraper_maps.py "Eiffel Tower" 5
{
  "query": "Joe's Pizza NYC",
  "source": "google_maps",
  "count": 3,
  "images": [
    {
      "url": "https://lh3.googleusercontent.com/...",
      "author": "Mary van Lutsenburg Maas",
      "author_profile_url": "https://www.google.com/maps/contrib/118270766331517021243"
    }
  ]
}

By default, the scraper prefers photos with visible author attribution and falls back to the place's official photo strip when fewer attributed photos are available — so you always get num_images results if any exist. Attributed photos are returned first, then official ones fill the remaining slots. Pass require_attribution: true to disable the fallback and only get attributed photos.

BeautifulSoup DuckDuckGo (no Selenium, fast)

uv run python image_scraper.py "cute cats" 5

Lightweight fallback. Only returns the proxied image URL and the alt text — no source attribution. Use this only when you need a fast scrape and don't need to credit the result.

Specifying the number of images

All scrapers accept an optional second argument for the result count (1–50, default 5):

uv run python image_scraper_selenium.py "sports cars" 10
uv run python image_scraper_maps.py "Eiffel Tower" 8
uv run python image_scraper.py "golden retriever puppy" 5

MCP Server

image_scraper_mcp_server.py exposes two tools:

Tool

Use for

Returns

search_images

General image search (DDG)

Text output with Credit:, Title:, Dimensions:, Alt: per result

search_maps_images

Place photos (Google Maps)

JSON output with author and author_profile_url per photo

search_images parameters

  • query (required): Search query

  • num_images (optional, default 5, range 1–50): Number of results

  • headless (optional, default true): Run Chrome headless

search_maps_images parameters

  • query (required): Place name (e.g. "Eiffel Tower", "Joe's Pizza NYC")

  • num_images (optional, default 5, range 1–50): Number of photos

  • headless (optional, default true): Run Chrome headless

  • require_attribution (optional, default false): If true, only return photos with visible author attribution. If false (default), prefer attributed photos but fall back to the place's official photo strip when fewer attributed photos are available.

See MCP_SETUP.md for installing the server as a Docker-based MCP tool for Claude Desktop or other MCP clients.

Tests

.venv/bin/python -m unittest discover -s tests

87 tests cover both scrapers and the MCP server, with all browser interactions stubbed so they run in seconds without Chrome or network access:

  • tests/test_image_scraper_ddg.py — DDG URL decoding, dimensions parser, tile attribution, MCP text formatting

  • tests/test_image_scraper_maps.py — Google Maps review-photo extraction, fallback, MCP JSON formatting

Notes

  • DuckDuckGo is generally more scraper-friendly than Google.

  • The Selenium DuckDuckGo scraper is more reliable than the BeautifulSoup version and returns proper attribution.

  • Some images may be low resolution or thumbnails; check width and height before reuse.

  • Respect robots.txt and the terms of service of DuckDuckGo and Google Maps.

  • Reuse of copyrighted images may require permission from the publisher; the attribution fields tell you who to ask.

License

MIT License — feel free to use and modify as needed.

Available Tools

2 tools
search_imagesA

Search for images on DuckDuckGo and return image URLs with full source attribution (source page URL, publisher domain, article title, image dimensions). Each result carries credit info so callers can attribute the image instead of just seeing a proxied redirect URL. Uses Selenium to scrape image results from DuckDuckGo Images.

ParametersJSON Schema
NameRequiredDescriptionDefault
queryYesSearch query for images (e.g., 'cute cats', 'mountain sunset')
headlessNoRun browser in headless mode (default: true)
num_imagesNoNumber of images to retrieve (default: 5)

TDQS

A4.2/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations provided, so description carries full burden. It discloses that Selenium is used (browser-based scraping, may be slow), and that each result includes source page URL, publisher domain, article title, and image dimensions. This is adequate transparency.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Three sentences, no fluff, front-loaded with the main purpose. Every sentence adds value.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

No output schema, but description explains return type (URLs with full attribution). Covers source (DuckDuckGo), method (Selenium), and output details. Given tool simplicity, it is sufficiently complete.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so baseline is 3. Description does not add significant parameter-specific meaning beyond the schema. The query example is helpful but minimal. Parameters are well-documented in schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

Description clearly states 'Search for images on DuckDuckGo and return image URLs with full source attribution'. The verb 'search' and resource 'images' are specific, and the sibling tool 'search_maps_images' indicates a different domain (maps), so there is no ambiguity.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Description highlights that the tool provides attribution info, distinguishing it from proxied redirect URLs. It implies this should be used when attribution is needed, but does not explicitly state when not to use it or mention alternatives beyond the sibling context.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

search_maps_imagesA

Search Google Maps for a place and return photo URLs from the place page. Accepts a place name like 'Eiffel Tower' or 'Joe's Pizza NYC'. Uses Selenium to scrape photos from the Google Maps place panel — no API key required. Returns a JSON object with each photo's URL, author, and author profile URL so callers can credit the photographer if they reuse the image.

ParametersJSON Schema
NameRequiredDescriptionDefault
queryYesPlace to search for on Google Maps (e.g., 'Eiffel Tower', 'Joe's Pizza NYC')
headlessNoRun browser in headless mode (default: true)
num_imagesNoNumber of photos to retrieve (default: 5)
require_attributionNoIf true, return only photos with visible author attribution. If false (default), prefer attributed photos but fall back to the place's official photo strip when fewer attributed photos are available. Attributed photos are always returned first.

TDQS

A4.5/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description fully discloses behavior: it uses Selenium for scraping, no API key needed, and returns photo URLs with author info. It also explains the require_attribution fallback logic. It lacks mention of potential failure cases or rate limits, but is transparent about the core scraping behavior.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is three sentences, each efficiently adding information: purpose, examples, and method with return structure. It is front-loaded with the core action and avoids unnecessary words.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

The description covers the return value (JSON with URL, author, author profile) and all parameters. It mentions the scraping method and lack of API key. However, it does not discuss potential issues like missing photos for some places or performance constraints, which would make it more complete for a tool with no output schema.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, establishing a baseline of 3. The description adds significant value beyond the schema: it explains the query parameter with examples, describes headless mode behavior, and gives a detailed explanation of require_attribution's fallback logic and ordering, which the schema description lacks.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool searches Google Maps for a place and returns photo URLs with attribution info. It distinguishes itself from the sibling tool 'search_images' by specifying it is for Google Maps place photos, not general image search.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides good usage context: accepts place names, uses Selenium (scraping), no API key required, and returns photos with attribution for crediting. However, it does not explicitly say when to use this tool versus the sibling 'search_images', which would further clarify selection.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.

  1. 2 tool updatesv0.1.0
    • First observedsearch_images
    • First observedsearch_maps_images
Install Server
RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

TDQS

A4.1/5.0
Disambiguation5/5

Both tools search for images but from distinct sources (DuckDuckGo vs Google Maps), with clear descriptions that eliminate confusion.

Naming Consistency5/5

Both tool names follow a consistent 'verb_noun' pattern ('search_images' and 'search_maps_images'), with no mixing of conventions.

Tool Count3/5

With only two tools, the server feels minimal for an 'Image Scraper' scope, though it may be intentionally focused on two specific sources.

Completeness2/5

The tool set lacks coverage for common image search sources (e.g., Google, Bing) and offers no download or metadata extraction tools, leaving significant gaps.

Maintenance

ActivitySlowing
ResponsivenessSyncing

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Related MCP Servers

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/spoiledbeef/image-scraper-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server