What is the Hybrid Edition intent router?

A dual-process router that sends each user query to the right worker agent: Jev makes the fast typed decision, and Gemma reviews only the cases that need deliberation.

What does Jev return in this design?

One System One request returns four things: the intent as a Choice, a human-review probability as a Noul, the utterance type as a Choice, and a Score-based guardrail risk from 0 to 4.

When does Gemma get invoked?

Only when System 1 is uncertain or the guardrail score is elevated; its job is to review Jev's full read and confirm, override or escalate to a human rather than answer every query.

What are the fast-path thresholds?

Confidence at least 0.6, noul at most 0.5 and guardrail score below 1.5 routes straight to a worker with no language-model cost; scores of 3.0 or more block immediately to a fallback worker.

Why does the README call this economical?

The point is that the expensive, slow reasoner runs only on queries that need deliberation, while everything else resolves as a cheap typed decision, typically in 70-500 ms.

Does it need API keys?

The digest does not say how Jev or Gemma are hosted or authenticated; no key requirements, install steps or dependencies are visible in the captured excerpt or metadata.

What is the project's status?

Python, 0 stars and 0 forks, created and last pushed 2026-10-02 with no license listed, and the README excerpt stops inside the routing-policy diagram.

Jev directory

CommunityJev Integrations & Adapters

Intent Router — Hybrid Edition

A dual-process intent router: Jev acts as System 1, returning the intent, human-review probability, utterance type and a score-based guardrail risk in 70-500 ms, while Gemma acts as System 2 only when Jev is uncertain or risk is elevated.

See highlights Open github.com

Stars: 0
Language: Python
Reading time: 2 min
Published by: Community
Author: anujgupta9860
Last verified: Oct 2026

anujgupta9860/intent-router-hybrid

User repository on GitHub

View on GitHub

Hybrid intent router: Jev as System 1 (Choice+Noul+Score guardrail) and Gemma as System 2 reviewer, dual-process routing to A2A worker agents

Forks: 0
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.

Tagscommunity github python routing agents

Works forJev for Routing Jev for Decision gating

Used inJev for Software & DevTools

Languages: Python
Capabilities: Open source

Highlights

One Jev request returns intent (Choice), human-review probability (Noul), utterance type (Choice) and a 0-4 guardrail Score.
Fast path requires confidence >= 0.6, noul <= 0.5 and score < 1.5; those queries skip the language model entirely.
Guardrail scores of 3.0 or higher block and route to a fallback worker without consulting System 2.
Gemma reviews only uncertain or elevated-risk cases and can confirm, override or escalate to a human.
Typical System 1 latency is 70-500 ms; 0 stars and no license, created and pushed 2026-10-02.

Watch out

README and GitHub metadata reviewed 2026-10-17; the captured text is cut off inside the routing-policy diagram, so the worker-agent side, escalation handling, setup steps and whether Gemma runs locally or over an API are not visible.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

Facebook: Intake Router keeps diagnosis out of the answer space at the front desk.

FacebookClinicnotes

X: OpenRouter Decision Model Rankings

X: herdr + jev, an agent router

X: A group chat bot with decision room routing

From the Jev guides

Original write-ups on what this entry is for.

Decision gates in agent frameworks: what each integration ships

Oct 16, 2026

How Pydantic AI, Mastra, TanStack AI, DSPy, Spring AI, and LiteLLM expose Jev: where the typed decision sits in the loop, what each integration returns, and which big frameworks are only adjacent.

Jev Integrations & Adapters#agents#frameworks#integrations

Read article

From the Jev AI model community

Posts from builders shipping with Jev right now.

Follow @typesafeai