Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

typesafe-ai-java

Community JVM SDK for System One with Java core, Kotlin DSL, Spring Boot and Spring AI; works with Jev and LAYA.

Librarybeta
View on GitHub ↗

Overview

typesafe-ai-java is a community-maintained SDK that speaks the System One wire contract from the JVM. It works with hosted Jev, Jev via Vercel AI Gateway or OpenRouter, self-hosted LAYA through laya-serve, and open-weight servers that implement the Jev HTTP API such as openjev-sglang. Models are discovered at runtime with listModels.

It ships four Maven Central artifacts: a Jackson-only core with typed questions, answers, retries and errors; an idiomatic Kotlin DSL with confidence helpers; a Spring Boot starter that auto-configures a client bean; and a Spring AI module that exposes any System One model as a ChatModel or a prompt-guard advisor. The latest release noted is 0.2.3. It requires Java 25. It is not an official TypeSafe product.

What the decision model decides

TypeQuestionAnswers
ChoicePick one option from a list—
Yes / noIs this statement true?yes · no
Yes / noShould this prompt be blocked (Spring AI prompt-guard advisor)?(inferred)yes · no

How it works

  1. Add the core, Kotlin, Spring Boot or Spring AI artifact.
  2. Point the client at Jev, a gateway or a self-hosted server.
  3. Send state with typed questions.
  4. Branch on the returned probabilities in code.

Strengths

  • One SDK for hosted Jev, gateways and self-hosted LAYA or OpenJev servers.
  • Spring Boot and Spring AI integration.
  • Published to Maven Central.

Caveats

  • Requires Java 25.
  • Unofficial community SDK.

Getting started

Add com.jamilxt:typesafe-ai-java-core:0.2.3 to your Maven or Gradle build.

At a glance

Type
Library
Status
beta
Language
Java, Kotlin
Runtime
JVM library, Spring Boot
Has code
yes

Use cases and integrations

How curated lists describe it

  • Community-maintained Java SDK for the TypeSafe AI System One (Jev) API. Not an official TypeSafe product.

    hellogumbo/awesome-jev

Frequently asked questions

Which backends does it support?

Hosted Jev, Jev via Vercel AI Gateway or OpenRouter, LAYA via laya-serve, and OpenJev-compatible servers.

Does it integrate with Spring?

Yes, via a Boot starter and a Spring AI ChatModel and advisor module.

Is it official?

No, it is community-maintained.

Related projects

  • laya-php

    Laravel-ready PHP SDK for self-hosted Laya that returns typed decisions as PHP enums, ints and bools, with a test fake.

    LibrarybetaPHP
  • Ax

    Ax, a DSPy-style multi-language LLM framework, listed with a native TypeSafe provider for Boolean, Choice, Score and raw Jev questions.

    PluginproductionTypeScript
  • Effect

    Effect TypeScript library whose @effect/ai-typesafe package maps classify, probability and rating operations to Jev questions.

    ProjectproductionTypeScript
  • instructor-php

    PHP monorepo whose Polyglot module adds a TypeSafe driver for typed Noul, Choice and Score decisions alongside LLM inference.

    ProjectproductionPHP
  • req\_llm

    Elixir library for unified LLM calls across providers, with an evaluate interface that calls Jev via TypeSafe or OpenRouter.

    ProjectproductionElixir
  • synkora-ai

    Self-hosted multi-tenant AI agent platform listed as including optional TypeSafe client tools for classification, scoring and yes/no.

    LibraryproductionPython