Systematic matching across the databases that were never built to connect. Surfacing leads hidden by jurisdictional gaps.
NamUs. Doe Network. Charley Project. State and county records. Missing persons and unidentified remains are logged across dozens of fragmented systems with no automatic cross-matching. RECALL builds the layer that connects them — scoring every plausible pair, validating the top candidates, and referring leads to the professionals who can confirm them.
December 1983. A fishing boat was found drifting in San Francisco Bay with its engine still running. No one was aboard. 76 days later, unidentified remains were recovered 5.4 miles south. In 40+ years, no one ever compared them.
RECALL is built by Rob, an independent researcher developing AI-assisted matching algorithms to help identify unidentified remains and bring missing persons cases to resolution.
This is not law enforcement. This is a triage layer — systematically surfacing leads that should be evaluated but never have been. Validated leads are referred to Medical Examiner offices for formal comparison.
Every algorithm version, every decision, every result is documented publicly. The methodology should be transparent enough for others to replicate and improve.
RECALL has flagged dozens of potential matches across California, Illinois, and beyond. Subscribe to follow the investigations.
Subscribe on Substack →