NewBeamWeaver
The example shows both direct multi-type classification and Jev as a router classifier, with an explicit minimum-confidence setting before dispatching to a selected model.
Added to Jevfast
NewHow it uses Jev
BeamWeaver's example invokes Jev on a support ticket with one Noul urgency question, one Choice team-routing question, and one Score severity question. The answers are printed with model, usage, cost, latency, and request ID; the Jev model is also installed as a classifier in a dynamic model router, where confidence gates routing between two chat models.
What Jev decides
Support ticket urgency and team routingExample answers · not a recorded Jev response · Source ↗
Question 1 · urgent
YOUR APP
INSTRUCTION
Does ticket describe an urgent operational disruption?
STATE
Ticket text: checkout is down and customers cannot place orders.
JEV · NOUL
YesNo
Question 2 · team
YOUR APP
INSTRUCTION
Which team should investigate the ticket?
STATE
Ticket text: checkout is down and customers cannot place orders.
JEV · CHOICE
- engineering
- billing
App workflow
- Invoke Jev
Submit the ticket and three named typed questions to the decision model.
- Inspect decision telemetry
Print typed answers and request accounting fields.
- Route model
Use Jev as a classifier with a configured 0.8 minimum confidence to select a chat model.
Why it is interesting
The example shows both direct multi-type classification and Jev as a router classifier, with an explicit minimum-confidence setting before dispatching to a selected model.