Do I need the Python implementation to run J++ programs?

No. You build with Rust 1.85 or newer via cargo install from rust/crates/jpp, and the installed binary runs without Python or Cargo; the Python sources remain only as a behavior reference.

How is this different from just prompting a model?

It makes questions, methods, and their compositions reusable values, so a program can build its next question or accept and return methods; prompting alone does not give you that composition.

Are the example runs free or do they call a live judge?

The shipped examples run offline on recorded answers at no cost, while purpose-only live runs described in progress notes needed over a thousand calls; check which mode an example uses.

What can J++ not do yet that I should plan around?

Reusable source-library methods, consistent composition rules, and an application built on standalone source are listed as next steps, so expect sharp edges and changing semantics.

What do the demos show, and are the numbers from real runs?

They live on jpp.towow.net/demos; cards 01-04 replay real runs and every number comes from that run's report and ledger, while the pages state that none of the cases is fully settled. Card 06 replays a live speaker-coach run where JEV answers two merged multiple-choice questions about every 1.3 seconds.

What did the real-machine runs cost?

A purpose-only entry ran on 2026-10-01 over 400 companies and one resume: 1,634 calls for $0.0857 with no hand-written question text. A GitHub issue-triage run cost $0.189 for its final arm and agreed with the maintainers' labels 69.00% of the time, against 71.75% for the project's own code.

What does CI guarantee about the docs and examples?

CI installs jpp on clean Ubuntu and macOS machines and runs the five examples with the installed binary, and it runs every code block in the docs, comparing the output pasted beneath each, on every push; if any fails, CI is red. The five examples run offline on recorded answers at no cost.

CommunityJev Integrations & Adapters

Jpp

Experimental language where questions and methods are composable values, with standalone .jpp sources run by a native Rust kernel and a Python implementation kept as reference.

See highlights

Open github.com

Stars: 30
Language: Rust
License: MIT
Published by: Community
Author: Towow-ai
Last verified: Oct 2026

Towow-ai/jpp

Organization repository on GitHub

View on GitHub

J++: an experimental language with standalone source and a Rust runtime. Compose questions and methods. 独立源码，组合问题与方法。

Forks: 3
Last pushed: Updated Oct 2026

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

About this project

Compose questions. Compose methods. Compose the compositions. 简体中文 · Why J++ · Ecosystem reassessment · Progress · Language design & grammar · Contributing. J++ is an experimental programming-language project exploring semantic judgment as a programmable operation. Questions are values. Methods are values. A composed method can become a building block in another method.

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

Tagscommunity jevbest catalog-import

Works forJev for Classification

Used inJev for Software & DevTools

Languages: Python

Highlights

Write standalone .jpp sources; the Rust kernel parses, checks rules, and executes via one core.
Installs with cargo install from rust/crates/jpp and needs Rust 1.85 or newer to build.
Ships five examples covering purpose runs, changing processes, modules, and ledger replay.
Installed binary runs without Python or Cargo; examples run offline on recorded answers.

Watch out

Experimental language project; reusable library methods and composition rules are still roadmap items. 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

jev-tree

Typed Python planner runtime that turns batched Jev action distributions into executable probability trees and graphs for finite decision tasks.