TypeSafe Jev · the router in front of your agents

AI request routing — send every request to the right team or agent with a priority attached

Paste a request or ticket and Jev returns the owning team, a priority tier, and whether it needs a human — a typed routing decision with calibrated confidence, ideal as the switchboard in front of an agent stack.

Run request / agent routing now

This tool is pre-configured for request / agent routing. Paste your text and Jev returns the typed decision below — the same call the showcase example was captured from.

0 chars
Your typed decision appears hereCategory · yes/no · score, each with confidence
Typed decision + calibrated confidence — nothing to parse, nothing to hallucinate.

A real request / agent routing decision

Captured live from this exact tool — the text that was pasted, the typed decision Jev returned, and the measured latency and cost. Nothing here is mocked up.

Request / agent routing492 ms

Our checkout page throws a 500 on Safari after the last deploy — customers cannot pay. Need this looked at ASAP.

TeamEngineering100%
Engineering100%
Ops0%
Sales0%
Product0%
PriorityHigh · 2.00100%
Needs HumanYes · 93%
typesafe/jev-1.13-20260917$0.000020

How request / agent routing works

01

Paste the incoming request

A ticket, an intent, a natural-language ask. Jev reads it as-is — no intent taxonomy to maintain and no slot-filling.

02

Jev picks a team, a priority, and a human flag

A single pass returns the destination (engineering / product / design / ops / sales), a priority tier, and a yes/no on whether it is safe to automate or must go to a person.

03

Wire it as your routing layer

Feed the chosen team and the needs-human flag into your dispatcher: auto-route confident, low-risk requests to the right agent and hold the rest for a human — all on one sub-500 ms call.

Request / agent routing — frequently asked

Why route with Jev instead of an intent classifier plus if/else?

An intent classifier gives you a label; you still hand-code the priority and the human-vs-bot call. Jev answers all three in one pass and attaches confidence to each, so your dispatcher logic is a few thresholds instead of a growing rules file.

How does the needs-a-human flag help an agent pipeline?

It is the safety valve in front of autonomous agents: high-confidence, automatable requests flow straight to the right agent, while anything Jev marks as needing a human is parked for review — so you automate aggressively without automating the risky calls.

Can I change the set of teams it routes to?

Yes. The teams are typed criteria, not a fixed enum baked into a model. Rename them, add a queue, or collapse two — the routing decision adapts without retraining.

What does the priority score actually represent?

It is an ordered tier (low / medium / high) returned as a number plus a legend, with per-tier probabilities. You can sort a queue by the raw score or bucket it — the calibration is yours to use.

Is it fast enough to sit in the request path?

The routing sample on this page returned in 492 ms end to end. That keeps a synchronous routing hop well inside a normal request budget, so you can decide before you dispatch.

Put a decision in front of your dispatcher

Paste a real request and see the team, the priority, and the human flag Jev would hand your router — in a single call.

Route a request