10 actions in 9 pieces of software · ranked by words, meaning and measured trust · 635 ms
modelrecord
- barakplasma/jev-omni
Jev-Omni (#1 on Image JevBench v0.1): ask a typed yes/no or multiple-choice question about an image, get a calibrated probability per option from one forward pass. Pinned akhilaaa3/Jev-Omni revision, Apache-2.0.
- barakplasma/jev-omni
- 2jevfound in code, not yet checkedno passport yet
- jev_ask
Ask several independent questions about the same state in ONE request. Jev prefills the state once and scores every question in a single forward pass, so extra questions add almost no latency.
- jev_ask
- 3jev-unrealfound in code, not yet checkedno passport yet
- jev_decide
Send explicit state and 1–32 Choice/Score/Noul questions to Jev. No tools are executed.
- jev_decide
- 4jev-usefound in code, not yet checkedno passport yet
MCP servershitianfang/jev-use
- jev_judge
Hand a batch of quick judgment questions to Jev (TypeSafe AI's System One model):
- jev_judge
- 5jev-browserfound in code, not yet checkedno passport yet
- jev_stop
Stops every running Jev task in the browser.
- jev_stop
- 6jevfound in code, not yet checkedno passport yet
- jev_ask
- jev_score
- 7jevjobfound in code, not yet checkedno passport yet
MCP serveradnanrules/jevjob
- status
The current pool, the search behind it, and whether Jev is configured.
- status
- 8Moreno.Jevfound in code, not yet checkedno passport yet
- jev_evaluate
Evaluate text or structured state with Jev typed questions; submitted state is sent to TypeSafe.
- jev_evaluate
- 9jev-mcpfound in code, not yet checkedno passport yet
- jev_status
Inspect a live Jev browser session without taking another browser action.
- jev_status
Labels are the instruments' verdict and are never changed by ranking. Agents and robots get this same answer as JSON at /api/v1/find?q=…&answer=1 and through the MCP server.