Documentation

Find Any Photo is a remote MCP connector and a website. It gives Claude tools that search real public-domain photographs by what they show, across Library of Congress collections, US government photographers and NASA, and return each photo with the reason it is public domain and a credit line. Nothing is installed on your computer, and nothing is AI-generated.

Set up

Connector address: https://findanyphoto.com/api/mcp. You sign in once with an email code or a Google account.

• Claude (web, desktop, iPhone, Android): add it in one click, or Customize → Connectors → Add custom connector.
• Claude Code: claude mcp add --transport http find-any-photo https://findanyphoto.com/api/mcp, then /mcp to sign in.
• Cursor and other MCP apps: add {"url": "https://findanyphoto.com/api/mcp"} to your MCP settings.

Tools

search_photos (query, decade, source, photographs_only, safest_only, limit (all but query optional))

Finds the photos that match a description and returns each with its photo_id, date, a preview image, the archive, the licence with the reason it is public domain, and the photo's page. One search per photo you need; each counts once against your monthly searches.

get_photo (photo_id)

Shows a photo's full archive record (description, photographer, collection, size), the reason it is public domain and the record that was read from, the credit line, and a link to download the full-size file. Free.

credits (photo_ids)

Writes the credit lines for the photos you used, with the reason each is public domain, ready for a video description or caption. Free.

my_usage (none)

Shows the signed-in account, its plan, searches used this month and the size of the index.

whoami (none)

Shows which account the connector is signed in with. Useful to check the connection.

Example prompts

• Here's my script about the Great Depression. Find a public-domain photo for each paragraph and make a list with credits.
• Find a photo of the Wright brothers' first flight, safest licences only, and give me the download link.
• I need photos of New York street life in the 1900s for a thumbnail. Photographs only, no prints.
• Give me the credit lines for photos 4021, 877 and 15530.

Sources and limits

• Sources today: Library of Congress collections and US government photographs through their copies on Wikimedia Commons, and NASA's Image and Video Library. More are being added; see licences.
• The index holds historical photos and NASA photos. Modern everyday scenes are mostly still under copyright and aren't in it; Carol M. Highsmith's photos of America (1980s to today) are the main exception.
• Free plan: 50 searches a month with an account (5 a day on the website without one). Pro: 2,000 a month (pricing). Claude and the website share the same count. Photo details, downloads and credits never count.
• Matching reads the archives' own catalogue records, which can be brief or old-fashioned. Look at each photo before you use it.

Privacy

The connector receives only what Claude sends to a tool: your search text and the photos you pick. It doesn't read your conversations, memory or files. Full details: privacy policy.

Troubleshooting

• No good match: describe the picture (who, doing what, where, when) rather than a single word, and drop filters.
• A photo is a print or a poster: turn on Photographs only (photographs_only in Claude).
• Pro doesn't show in Claude: ask Claude to run my_usage and check the email matches the one on your account page.
• Anything else: support@wikivec.com.