JevMade Sign in
← Back to experiments

Apps & data pipelines

JEV Dispatch

Bookmark: JEV Dispatch Keep this in your collection.
Leave a noteWhat would you try with this? : JEV Dispatch

Only you can see your notes.

Source screenshot of JEV Dispatch
SOURCE SCREENSHOTFull screenshot ↗

What it does

Jev dispatches taxis through a simulation of Baku's streets.

How you can use it

Start by collecting sample ride requests and listing your dispatch questions. You might ask the system to pick the nearest driver, rate trip urgency, and estimate whether a request looks fake. Then decide when people step in: dispatch obvious trips immediately, block clear fakes, and route doubtful bookings to a person for review.

A developer can connect this decision model to your dispatch software using an access key from TypeSafe. Because automated fraud estimates can make mistakes, set your limits so that uncertain orders pause for a person to check rather than stranding genuine riders.

Maker-reported (not independently measured by JevMade): median 361 ms and p95 1,168 ms per Jev decision, against 983 ms and 1,592 ms for Gemini 3.5 Flash Lite (author-reported) · about $0.036 per 1,000 decisions against about $0.40 for the LLM (author-reported) · 73-86% agreement with the LLM on the same driver (author-reported)

Primitives
choice, score, noul
Platform
JavaScript
Added
Project created
GitHub stars
0 (snapshot, not live)

Keep this for later

Sign in to bookmark experiments, guides and videos, and keep notes only you can see.

Continue to sign in

We’ll bring you back to this listing.