How these statuses are determined
Each company's status is researched by an internet-capable language model that is instructed to find and cite real, linkable sources for its claim. The raw answer is then run through an automated verification step before anything is published.
Verification rules
- A status of active, acquired, defunct, etc. is only published when it is backed by at least one well-formed source URL. Claims with no usable source are withheld and shown as unverified.
- Source URLs that are malformed or fabricated are discarded before grading — a "trust me" citation counts for nothing.
- Terminal claims (a company being defunct, bankrupt, or acquired) are held to a higher bar, because wrongly declaring a living company dead is the most damaging error this kind of tracker can make.
- Acquisition / merger / rename claims must name the successor entity.
- The model's self-reported confidence is an input, never the sole basis: high confidence with no usable source still fails verification.
Caching
Results are cached locally with a time-to-live (corporate status changes over months, not seconds), so repeat builds are cheap and deterministic and the underlying research API is not hit needlessly.
Limitations
Automated research can still be wrong or out of date. Every published claim links its sources so you can check them yourself. Treat this site as a starting point, not a system of record.