Skip to content

J++

Experimental language where questions and methods are composable values, with a Rust runtime that speaks /v1/systemone and replays runs.

Serviceresearch

Overview

J++ is an experimental programming language exploring semantic judgment as a programmable operation. Questions are values, methods are values, and composed methods can become building blocks in other methods. Standalone .jpp source runs on a native Rust implementation that parses, checks language rules and executes methods through one shared kernel; an earlier embedded Python 3.12 implementation remains as a behaviour reference. Both speak the /v1/systemone format.

Examples cover three ways in: a batch of items from one sentence of purpose, a process that keeps changing in a small world, and modules you fill in yourself, plus filling missing information and replaying a run from its ledger. All five examples run offline on recorded answers at no cost, and CI installs the binary on clean Ubuntu and macOS machines and checks every documented code block's output. A demos page replays real runs and states what is unsettled. Documentation is in English and Chinese. The roadmap lists reusable source libraries and consistent composition rules as next steps.

What the decision model decides

TypeQuestionAnswers
OtherUser-composed questions and methods evaluated by a System One model—

How it works

  1. Write a .jpp program composing questions and methods
  2. Run it with the jpp command
  3. Kernel sends questions to a System One endpoint or replays recorded answers
  4. Results and a ledger are produced for replay

Strengths

  • Offline examples via recorded answers
  • CI verifies documentation output
  • Native binary needs no Python at runtime

Caveats

  • Experimental language with unsettled design
  • Rust 1.85+ needed to build
  • Limited library of methods so far

Getting started

git clone https://github.com/towow-ai/jpp.git && cd jpp/rust && cargo install --locked --path crates/jpp && jpp run examples/composition.jpp

At a glance

Type
Service
Status
research
Language
Rust
Runtime
native CLI interpreter (Python reference implementation)
Has code
yes
Has tests
yes

Use cases and integrations

How curated lists describe it

  • Experimental language in which a question is a value and methods compose them, with a Python foundation and a Rust runtime, both speaking `/v1/systemone`. Every demo replays a recorded run and says what is unsettled. Chinese and English.

    MrJev/awesome-jev

Frequently asked questions

What is a question value?

A typed judgment that can be passed around and composed into methods like any other value.

Can I try it without a key?

Yes, the examples replay recorded answers offline.

Which implementation should I use?

The native Rust implementation; the Python one is kept as a behaviour reference.

Related projects

  • Self-hosted Jev-compatible /v1/systemone server using an encoder with decision heads for parallel noul, choice and score answers.

    ServicebetaPython
  • Simon Willison's LLM plugin that adds Jev as a model for yes/no, choice and score questions from the command line.

    PluginproductionPython
  • Agent Zero plugin adding a typesafe_query tool that asks Jev Choice, Noul and Score questions and shows probability cards in chat.

    ProjectbetaPython
  • Go integration bringing TypeSafe Choice, Score and Noul to Google's adk-go agents, with a typed client, routing agent and callbacks.

    ProjectbetaGo
  • Chinese everyday-dilemma app where an LLM lays out the options and Jev picks one, revealed after you record your own leaning.

    Projectbeta
  • Docker service exposing local Laya, Von, a CLM head or hosted Jev through one TypeSafe-compatible HTTP and MCP endpoint.

    MCP serverbetaPython