What it does
Instruction sets are versioned JSON, so one dataset can be replayed through every model. Jev, Von and Laya are all reached through the same endpoint contract, so only the wire translation differs.
Apps & data pipelines
A self-hosted console that runs VirusTotal summaries through System One models and evaluates those models against each other on the same labelled dataset.
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
Instruction sets are versioned JSON, so one dataset can be replayed through every model. Jev, Von and Laya are all reached through the same endpoint contract, so only the wire translation differs.
For a security research project, use IOCArena to compare how AI systems judge the same known examples. A developer can set up the app and feed it internet addresses, domain names, or file hashes: digital fingerprints of files, not the files themselves. It gets security reports from VirusTotal, then asks each AI to judge them.
VirusTotal needs its own software access key. Testing Jev also needs one from TypeSafe, the outside AI provider; the other two supported AI systems run locally. This is a comparison tool, not a protective shield. An AI label does not prove an address or file is safe.