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

new-api-typesafe-plugin

Task plugin for the new-api LLM gateway that adds a native /v1/systemone endpoint for Jev with auth, routing and token billing.

Pluginbeta
View on GitHub ↗

Overview

new-api-plugin-typesafe adds TypeSafe System One (Jev) support to QuantumNous/new-api, a self-hosted LLM gateway. Clients send state and typed questions to /v1/systemone on the gateway, which handles authentication, channel routing, model mapping, billing and logging, then forwards the request and returns TypeSafe's native response unchanged. It deliberately does not emulate chat-completions endpoints. The README is in Chinese.

Two upstreams are supported, detected from the channel base URL: TypeSafe directly with a TypeSafe key, or Vercel AI Gateway with an existing vck_ key, with the plugin translating noul to boolean and usage field names as needed. Model mapping lets clients use a short name like jev and upgrade versions by changing the mapping. Token-based billing uses usage expressions, and missing upstream usage is never fabricated. It installs from a marketplace index with sha256 checks or by manual upload.

What the decision model decides

TypeQuestionAnswers
OtherAny Choice, Score or Noul question passed through the gateway(inferred)—

How it works

  1. Install the plugin from the marketplace source or upload it.
  2. Create a Task Plugin channel pointing to TypeSafe or Vercel AI Gateway.
  3. Clients call /v1/systemone on the gateway with their gateway token.
  4. Gateway bills by input tokens and logs each call.

Strengths

  • Typed decisions behind the same gateway, auth and billing as chat models.
  • Version upgrades through model mapping only.
  • Works with Vercel AI Gateway keys.

Caveats

  • Requires a running new-api instance.
  • Documentation is in Chinese.

Getting started

In the new-api console, add the marketplace source https://raw.githubusercontent.com/FFatTiger/new-api-plugin-typesafe/main/index.json and install TypeSafe AI, then add a Task Plugin (61) channel.

At a glance

Type
Plugin
Status
beta
Language
JavaScript
Runtime
new-api gateway plugin
Has code
yes

Use cases and integrations

How curated lists describe it

  • LLM gateway: adds a native `/v1/systemone` endpoint to new-api so typed decisions sit behind the same gateway as chat models.

    yibie/awesome-jev

  • TypeSafe AI System One (Jev) task plugin for QuantumNous/new-api — native /v1/systemone, synchronous evaluation, token billing.

    hellogumbo/awesome-jev

Frequently asked questions

Does it expose a chat-completions endpoint for Jev?

No, it only serves the native /v1/systemone protocol.

Can I use a Vercel AI Gateway key?

Yes, set the base URL to ai-gateway.vercel.sh and map model names to typesafe-ai/jev.

How is usage billed?

By input and output tokens via a configurable usage expression.

Related projects

  • ai

    Hack Club's lightweight LLM proxy, listed for adding a Jev forwarding endpoint behind its auth, limits and usage logging.

    MCP serverproductionTypeScript
  • decision-gate

    npm library that rate-limits, spend-caps, caches and redacts every Jev request made in a loop.

    LibrarybetaJavaScript
  • jev-accounts-hub

    Go gateway that pools multiple TypeSafe accounts behind the native System One API with user keys, quotas and failover.

    ServicebetaGo
  • jevkit

    Rust CLI for Jev that lints typed question sets offline, then sends them and prints probabilities, labels and scores as JSON.

    CLIbetaRust
  • jevtok-ts

    Offline TypeScript token counter and request-cost estimator for Jev 1.13, a port of the reconstructed jevtok encoder.

    PluginbetaTypeScript
  • x-herald

    Chinese-documented LLM gateway with virtual model routing, protocol conversion and a transparent Jev System One proxy.

    ServicebetaTypeScript