When an autonomous video pipeline or editor needs background B-roll, commercial stock sites block headless requests behind Cloudflare challenges or require per-user API keys, while raw Internet Archive searches mix unverified mirrors without licence tags into the top results. On 2026-10-20 we benchmarked Video Grab (https://videograb.openkrill.app/mcp) on the query "ocean waves" against the Wikimedia Commons MediaWiki API, Internet Archive Advanced Search and Pexels Videos. In 2,335 ms, search_free_videos returned 5 CC0 1.0 .ogv clips by Mostafameraji with direct upload.wikimedia.org download URLs, Creative Commons deed links and preformatted attribution strings.
What Video Grab provides across 3 MCP tools
Video Grab runs at https://videograb.openkrill.app/mcp with no account and no API key. It exposes three read-only tools designed for openly licensed media discovery and compliant personal exports:
search_free_videossearches Wikimedia Commons and the Internet Archive (mediatype:movies) for openly licensed video files (limit1 to 5, optionalsourcesarray of"commons"and"archive"). It keeps only items whose metadata records a public-domain, CC0, CC BY or CC BY-SA licence and builds a ready-to-copyattributionstring for each clip.video_metadataaccepts a Wikimedia CommonsFile:title or URL, or an Internet Archive item identifier orarchive.org/details/URL, and returns the full file list (name,url,mime,bytes,kind), creator, licence deed URL, attribution line and any attached subtitle or caption tracks.own_video_exportreturns the official, terms-compliant steps and daily limits for exporting videos you uploaded yourself ontiktok,youtube,instagramorx, citing each platform's help documentation. It never rips third-party social videos or strips watermarks.
Step 1: Search CC0 ocean-waves videos and compare 3 rivals
We ran search_free_videos for "ocean waves" (limit: 5) on 2026-10-20 and compared it against three stock-video and archive sources:
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "search_free_videos",
"arguments": {
"query": "ocean waves",
"limit": 5
}
}
}
All 5 returned clips are CC0 1.0 videos on Wikimedia Commons by creator Mostafameraji (Ocean surface waves 06, 09, 05, 08 and 04), each with a direct https://upload.wikimedia.org/wikipedia/commons/... download URL and a complete attribution line such as:
"Ocean surface waves 06" by Mostafameraji, CC0 1.0 (https://creativecommons.org/publicdomain/zero/1.0/deed.en/), via Wikimedia Commons. https://commons.wikimedia.org/wiki/File:Ocean_surface_waves_06.ogv
| Tool or source | Wall time | Score | Results returned for "ocean waves" | Licence filtering and attribution readiness |
|---|---|---|---|---|
Video Grab (search_free_videos) |
2,335 ms | 9.6 / 10 | 5 CC0 1.0 .ogv videos (Ocean surface waves 06, 09, 05, 08, 04) |
Filters out unlicensed items; includes direct upload.wikimedia.org URL, deed URL and preformatted attribution line |
Wikimedia Commons API (w/api.php?action=query&generator=search) |
795 ms | 9.0 / 10 | Same 5 .ogv files |
Fastest raw response, but caller must construct iiextmetadata queries, strip HTML from Artist and assemble attribution strings |
Internet Archive (advancedsearch.php?q=ocean+waves+AND+mediatype:movies) |
1,222 ms | 6.0 / 10 | 5 items without licenseurl |
Top 5 hits were unrelated YouTube mirrors and music uploads with no licenseurl field and no direct file URL |
Pexels Videos (pexels.com/search/videos/ocean%20waves/) |
not run | not run | HTTP 403 challenge | Web search blocks automated fetches with a Cloudflare challenge; REST API requires an account and API key |
Step 2: Inspect a video file's metadata and download list
Pass any Wikimedia Commons File: reference or Internet Archive identifier to video_metadata to inspect its exact file streams and licence notes before downloading:
{
"jsonrpc": "2.0",
"id": 2,
"method": "tools/call",
"params": {
"name": "video_metadata",
"arguments": {
"ref": "File:Ocean surface waves 06.ogv"
}
}
}
The response returns title: "Ocean surface waves 06", creator: "Mostafameraji", licence: "CC0 1.0", download_url: "https://upload.wikimedia.org/wikipedia/commons/0/08/Ocean_surface_waves_06.ogv", mime: "application/ogg" and the files array.
Step 3: Look up official export steps for your own uploaded videos
When a creator asks an agent how to back up their own uploaded videos from YouTube, TikTok, Instagram or X, call own_video_export:
{
"jsonrpc": "2.0",
"id": 3,
"method": "tools/call",
"params": {
"name": "own_video_export",
"arguments": {
"platform": "youtube"
}
}
}
For platform: "youtube" (checked against YouTube Help answer 56100 and Google Takeout answer 3024190), the tool returns the five steps to download individual uploads from YouTube Studio (Content -> Menu -> Download as 720p or 360p MP4, up to five times per 24 hours) or export your full channel archive through Google Takeout.
Connect Video Grab to your AI client
- Video Grab MCP server reference and schema (
https://videograb.openkrill.app/mcp) - Video Grab Agent Skill (SKILL.md) for CC0 B-roll discovery and attribution checking
Sources
- Wikimedia Commons File:Ocean surface waves 06.ogv and Wikimedia Commons MediaWiki API (checked 2026-10-20)
- Internet Archive Advanced Search API (checked 2026-10-20)
- YouTube Help: Download YouTube videos that you've uploaded
Frequently asked questions
- Where can I find free CC0 or public-domain stock videos with ready-to-copy attribution lines?
- Call search_free_videos on https://videograb.openkrill.app/mcp with your search query. The tool queries Wikimedia Commons and the Internet Archive, filters out items without a clear public-domain, CC0, CC BY or CC BY-SA licence, and returns the direct file or item download URL, creator name, licence deed URL and a ready-to-paste attribution string.
- How do I get the direct download URL and license details for a video on Wikimedia Commons or the Internet Archive?
- Call video_metadata on https://videograb.openkrill.app/mcp with ref set to a Wikimedia Commons file title (such as File:Ocean surface waves 06.ogv), a Commons file URL, an Internet Archive item identifier or an archive.org/details/ URL. It returns the licence, creator, attribution text, direct video file URLs and any available caption files.
- How do I export or download my own uploaded videos from TikTok, YouTube, Instagram, or X?
- Call own_video_export on https://videograb.openkrill.app/mcp with platform set to tiktok, youtube, instagram or x. It returns the official, terms-compliant export steps and limits quoted from that platform's help center (such as YouTube Studio Content download and Google Takeout).
- Does Video Grab rip third-party videos from YouTube, TikTok or Instagram?
- No. Downloading other users' videos or stripping watermarks from TikTok, YouTube, Instagram or X violates those platforms' terms of service. Video Grab only searches openly licensed stock videos on Wikimedia Commons and the Internet Archive and provides official archive instructions for videos you uploaded yourself.