itsmostafa/system-one-connector

System One MCP connector to evaluate anything fast and cheap. Give your AI agent direct access to models like: Jev, D1, CLM and Laya

What it solves

AI agents often struggle with making quick, binary or categorical judgment calls because standard LLMs generate prose (e.g., "This seems urgent") rather than structured data. This makes it difficult for agents to branch their logic based on a confident answer versus a guess.

How it works

System One Connector provides a bridge between AI agents (such as Claude Code, Claude Desktop, Codex, Hermes, and pi) and "System One" models like Jev, CLM, Laya, and d1. These models are designed for judgment rather than text generation. Instead of a paragraph, they return a fixed format containing a probability or a specific option (e.g., a 0.95 probability of urgency).

Who it’s for

Developers building AI agents that need to perform classification, sentiment analysis, or urgency detection with high confidence and structured, machine-readable output that can be used directly in code logic.

Highlights

  • Structured Judgments: Supports three question types: yes/no (noul), single-choice options (choice), and scale ratings (score).
  • Multi-Model Support: Works with hosted models like Jev and d1, or self-hosted open models like CLM and Laya.
  • Multi-Question/Dataset Processing: Can handle multiple questions in a single call and process up to 500 records simultaneously.
  • Fast Performance: Jev typically responds in under half a second.
  • Easy Integration: One-command setup (evaluate setup mcp) for supported clients and a single static binary with no runtime dependencies.

Related

  • Project
    Anil-matcha/awesome-jev-by-typesafe

    A collection of use cases and patterns for Jev, a System One model by TypeSafe AI that provides typed decisions (Choice, Score, Noul) instead of generative text for software control flow.

  • Project
    wuyoscar/jev-skill

    A collection of specialized decision-making skills and workflows for coding agents, enabling them to classify, score, and choose actions using the Jev system.

  • Project
    razorback16/openjev

    An open-source decision server that provides fast, calibrated, and strictly typed answers (yes/no, choice, score) by reading model probabilities directly instead of parsing text.

  • Project
    BillionsBobby/JevRouter

    A fast agent routing layer that uses a specialized decision model to determine which tool or capability should be used next, reducing reasoning token waste and improving latency.

  • Project
    malevrigns/agent-jev

    A 0.6B parameter decision model that replaces generative text with direct probability distributions for agent routing, gating, and scoring, requiring zero decoded tokens.