What it does
In the maker's chess test, it solved 6 of 25 mate-in-one positions.
Benchmarks & research
jev-benchmark
Test Jev on chess puzzles and on figuring out which game character you're talking to.
Bookmark: jev-benchmark
Keep this in your collection.
Leave a noteWhat would you try with this?
: jev-benchmark
Only you can see your notes.
Recording unavailable. Open the experiment ↗
How you can use it
You can build game characters that know when a player is talking to them. A developer calculates basic facts first, like who is nearby. The app sends those facts and the player's words to TypeSafe. This outside AI service then decides which character should answer.
To start, your developer can adapt the project's speech testing rules. They will need an access key that connects the app to TypeSafe. The tool cannot calculate exact answers on its own. The app must feed it clear facts before asking it to make a choice.