Is that confidence a real Jev probability, and is it calibrated?

Unknown. The README says the pick comes with a confidence but never defines it, and reports no calibration measurement, so treat the number as an uncalibrated score until you measure it on your own task distribution.

What happens when two agents are plausible or none of them fit?

Not documented. There is no stated tie-break, no threshold below which the router abstains and no human-in-the-loop path, so the realistic worst case is a confident wrong pick with nothing to catch it.

Can I add my own agents, or are those five fixed?

Unclear. The README lists research, code, browser, support and writer as the handler set and says nothing about configuration, so whether the list is editable is a question for the source.

Does it hand the task to the chosen agent, or only tell me which one?

The README describes picking a handler and drawing a map. Dispatch is not mentioned, so the routing decision and the actual execution appear to be separate, and nothing here describes starting or supervising the chosen agent.

What are the setup commands for SwarmRouter?

npm i then npm run dev. The README documents no key, config file, or environment variable beyond those commands.

What language is SwarmRouter written in?

It is cataloged as TypeScript. The README names no framework, runtime version, or dependencies.

Is SwarmRouter part of a larger set of Jev demos?

Yes. It is one of eight demos cataloged under the GitHub user ndolinschi, alongside JevPlay, LaneBreak, McpMatch, PulseLane, SpendBrake, ToolGate, and TrustGate.

CommunityJev Demos & Experiments

Swarmrouter

SwarmRouter picks which agent or skill should handle a task and returns a confidence with the pick, drawn as a visual swarm map, with the choice made by TypeSafe Jev. Five handler types are named in the README.

See highlights

Open github.com

Stars: 0
Language: TypeScript
Published by: Community
Author: ndolinschi
Last verified: Oct 2026

ndolinschi/swarmrouter

User repository on GitHub

View on GitHub

Route tasks to research/code/browser/support/writer agents via TypeSafe Jev

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.Homepage

About this project

Pick which agent/skill handles a task (research / code / browser / support / writer) + confidence — visual swarm map powered by TypeSafe Jev.

Quoted from the project's README. Read the full README.

Tagscommunity jevbest catalog-import awesomejev

Works forJev for Classification Jev for Routing Jev for Tool calling

Used inJev for Software & DevTools

Languages: TypeScript

Highlights

Five handler types are named: research, code, browser, support, writer.
The output is a chosen handler plus a confidence value.
The result is presented as a visual swarm map rather than a plain list.

Watch out

The README is one sentence plus install commands. It does not define confidence — whether it is a Jev probability, a margin or a heuristic score — does not say whether the five handlers are configurable or hardcoded, and does not describe the low-confidence path: no threshold, no abstain, no human fallback. No routing accuracy, latency or cost figure is given, and nothing documents how the chosen handler is invoked afterwards, so routing and dispatch are most likely separate pieces. Copy reviewed against the repository README on 2026-10-04, not from running it.

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

Pi-Jev-Guide

Pi guard plugin that checks inputs and tool calls with local rules plus Jev semantic judgements, with redaction and verification reminders.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

zszz3/Pi-Jev-GuideTypeScript

flue-jev-demo

Standalone Flue example that routes support chat through Cloudflare Workers AI, using gpt-4o-mini for chat and Jev for typed intent and urgency routing.

Jev Demos & Experiments#community#awesomejev#catalog-import

Community

matthewp/flue-jev-demoTypeScript

ui-generator-instinct-jev

Demo that turns free-text UI descriptions into real shadcn components using only Jev typed Choice, Noul, and Score answers, with no generated code on the decision path.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

joevidev/ui-generator-instinct-jevTypeScript

jev-browser

TypeScript browser harness where an OpenRouter planner splits goals into subgoals, Jev picks each concrete action, OpenRouter fills text fields, and Playwright drives the page.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

vinilana/jev-browserTypeScript

jev-ra

Browser-use layer for CLI coding agents: Jev picks the operation and target element in one round trip per step, exposed over MCP with a CLI, a Python API and an HTTP server.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

brnyxx/jev-raPython

jev-lab

Two TypeScript monorepo products that keep the rules in the engine and ask Jev exactly one question: what should this person or this shogi side do next. Hundred runs a hundred NPCs with hunger, money, jobs, relationships and memories; Jev Shogi picks Gote's reply from engine-legal USI moves.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

jammaru/jev-labTypeScript

From the Jev guides

Original write-ups on what this entry is for.

Per-Tick Decisions in Games

Sep 28, 2026

A game tick is a state plus a legal action list. How Clash Royale, Minecraft, and the earlier Mario, Pokémon, StarCraft, and drone loops ask Jev one step at a time.

Jev Demos & Experiments#gaming#agents#control

Read article

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai