web-media-getter

web-media-getter is a Claude Code skill that runs one query across free image/video/GIF APIs and returns a single license-tagged list of results.

Search a pile of free media APIs with one query and get back a single, license-tagged list of images, videos, and GIFs — optionally downloaded with their attribution. It's one Python file, no dependencies to install, and it runs anywhere python3 does.

One query → five free sources fanned out at once, each result badged by source and license.

🤔 Why

When you need a real photo or a piece of archival footage — a hero image, a texture, 1930s factory film, a reaction GIF — you usually end up tab-hopping across NASA, Wikimedia, the Internet Archive, and a couple of stock sites, each with its own search box and its own licensing fine print. This collapses all of that into one command and hands the licensing back to you with every result.

📦 Install

You only need Python 3.8+ (already on most Macs and Linux machines). Grab the single script:

curl -O https://raw.githubusercontent.com/connerkward/web-media-getter-skill/main/webmedia.py
python3 webmedia.py "your query"

That's the whole install — no pip install, no virtualenv, nothing to keep updated.

🪟 On Windows

The script is pure standard library, so it runs on Windows too — you just need Python, which isn't preinstalled there. Easiest path (PowerShell):

winget install Python.Python.3      # one time, if you don't have Python
curl.exe -O https://raw.githubusercontent.com/connerkward/web-media-getter-skill/main/webmedia.py
python webmedia.py "your query"

Two Windows gotchas: the command is python (not python3), and use curl.exe explicitly so PowerShell doesn't intercept curl as an alias. Everything else — sources, downloads, attribution — works identically.

🤖 Using this with Claude Code or another AI agent? Install it as a skill and let the agent drive it — see docs/agents.md.

🚀 Use it

webmedia.py QUERY [--type image|video|gif] [--count N] [--source LIST|all|nokey]
                  [--download] [--out DIR]

# Images across every free source
python3 webmedia.py "apollo moon landing" --source all --count 8

# Historical footage from the Internet Archive
python3 webmedia.py "car factory 1930s" --type video --source internetarchive --count 5

# Reaction GIFs (needs a KLIPY or GIPHY key — see Sources)
python3 webmedia.py "thumbs up" --type gif --count 6

# Download the results and write attribution alongside them
python3 webmedia.py "saturn v" --source nasa --download --out ./downloads

• --source defaults to all; use nokey for the keyless sources only, or a comma list like wikimedia,pexels.
• --download saves each file into --out and drops an attribution.json next to them, so where each file came from travels with it.

Three media types, one interface

🗂️ Sources

The five no-key sources work the moment you run the script. The rest take a free API key — set the matching environment variable and they switch on automatically.

⚖️ Attribution & licensing

The tool reports a license for every result, but you check the terms before you use anything. A quick guide to what comes back:

• NASA — generally public domain.
• Wikimedia / Openverse — per-item CC or public-domain licenses; CC-BY and CC-BY-SA require credit.
• Internet Archive / Library of Congress — varies per item ("see item").
• Pexels / Pixabay — their own free licenses.
• GIF platforms — expect platform attribution in anything you ship.

--download always writes attribution.json so the provenance is recorded whether or not you read it at the time.

License

MIT © Conner K Ward. The tool is MIT; fetched media is licensed by its source — see above.

🧭 ckw-skills — part of Conner K. Ward's collection of Claude Code skills & MCP servers.