[ CASE FILE // internal tool ]Active
Datasink (sink-cli)
Data hygiene for music PR. Scrub, rinse, and soak contact lists.
An open-source CLI and browser engine to format, clean, deduplicate, and enrich contact spreadsheets with LLMs.
PR contact spreadsheets easily rot with bad email syntax, typos, and duplicate rows. Datasink cleans up these contact sheets: corrects UK-specific domain typos (like bbc.com -> bbc.co.uk), strips duplicates, and uses AI site scraping to find submission guidelines and outlet contacts.
Where it fits
Portfolio lead-gen and credibility builder. Sits top-of-funnel for technical developers.
Stack
Node.jsTypeScriptFirecrawl APIAnthropic SDK
[ ILLUSTRATION ]
$ sink wash contacts.csv --provider anthropic
[22:08:15] SCRUB: Validated RFC 5322 emails. Corrected: bbc.com -> bbc.co.uk
[22:08:16] RINSE: Jaro-Winkler Jaccard dedup (threshold=0.92) collapsed 47 duplicates
[22:08:19] STEEP: Crawled outlet URLs. Extracted presenter names and submission rules
Want a system like this built for your business? See consulting services →