Start
Repackage the Common Crawl records you care about into a new WARC archive in your own Hugging Face bucket. Sign in to run your own — or try a precomputed example right now, no account needed.
Sign in with Hugging Face to submit jobs to your account.
Examples — try without signing in
Each example runs on precomputed data: the form fills in, the job logs stream like a real run, and you get a working download link. Two use the structured fields, two use the advanced SQL filter.
Define your filter
Pick which records to keep. Use the guided fields for the common cases, or switch to raw SQL for full control over the columnar index.
Set at least one of the three below — registered domains, exact hostnames, or content languages. Each is optional on its own.
More options
Choose crawls
Select which Common Crawl archives to scan. Fewer crawls keep the index scan cheap.
Output destination
Where the repackaged WARC is written. Must be an existing Hugging Face bucket you can write to.
Advanced: hardware
Review & estimate
A cheap index-scan job counts the matching records and estimates the runtime and cost before you commit.
Estimating and repackaging run as HF Jobs on your own account (needs a Pro/Team plan). Sign in to continue — or use “Run on my own hardware” above to run it yourself (no account needed).
Run repackage
This launches the fetch job, streams progress, and writes the WARC to your bucket. You can cancel any time.