1. [Home](/)
2. [Products](/products)
3. Jev Judge MCP

Developer tooling · Open-source MCP server on PyPI Active system

# Jev Judge MCP. Independent verification for agents.

Jev Judge MCP is an open-source Model Context Protocol server built by PyModel LLC and powered by TypeSafe AI’s Jev model. Published on GitHub and PyPI, it gives coding and automation agents typed verify, classify, rerank, review, score, and gate tools that map calibrated probabilities to clear policy actions.

**Built for:** Engineers building MCP agent workflows, CI review gates, and applications requiring calibrated natural-language judgment.

[View Jev Judge MCP on GitHub →](https://github.com/PyModel/jev-judge-mcp) [View package on PyPI →](https://pypi.org/project/jev-judge-mcp/)

Calibrated judgment instead of self-grading

## Separate generation from verification inside MCP workflows.

Autonomous coding and operations agents routinely fail at two boundaries: rigid regex rules break on legitimate natural-language variation, while asking a frontier LLM to self-grade its own diff wastes tokens and rubber-stamps subtle omissions.

PyModel LLC built Jev Judge MCP (github.com/PyModel/jev-judge-mcp) to expose TypeSafe AI’s Jev judgment model as clean, typed MCP tools—verify, screen, find, classify, rerank, decide, compare, extract, review, gate, and score. Each call evaluates explicit evidence against a stated criterion and maps calibrated probabilities into deterministic policy decisions (auto, review, or escalate), giving Claude Code, Codex, Cursor, and custom pipelines an independent gate before acting.

## Typed judgment primitives in Jev Judge MCP

Open-source MCP server installable from PyPI (jev-judge-mcp) and GitHub.

### Verify, review & gate

Check whether a diff, document answer, or proposed action actually satisfies every stated acceptance criterion before committing or shipping.

### Classify, screen & decide

Route tickets, triage inputs, and moderate natural-language payloads using typed labels and calibrated confidence thresholds.

### Find, rerank & extract

Score retrieved passages for true relevance and extract structured fields without spinning up heavy frontier reasoning calls.

### Auto / review / escalate policies

Translate calibrated probabilities into repeatable operational boundaries so uncertain edge cases escalate cleanly to a human owner.

Verification pattern comparison

## Jev Judge MCP vs. frontier LLM self-grading vs. brittle regex rules

When to use calibrated small-model MCP judgment tools in agent and application pipelines.

__Comparison of Jev Judge MCP typed judgment tools against frontier LLM self-grading and static regex heuristics.__
| Evaluation Property     | Jev Judge MCP (TypeSafe AI Jev)                                                     | Frontier LLM Self-Grading Prompts                                | Regex / Keyword Heuristics                                            |
| ----------------------- | ----------------------------------------------------------------------------------- | ---------------------------------------------------------------- | --------------------------------------------------------------------- |
| Independence            | Separate evaluator model checking evidence against explicit criteria                | Same authoring context grading its own reasoning and assumptions | Independent of LLM, but blind to semantic meaning                     |
| Output contract         | Typed labels, calibrated probabilities, and auto / review / escalate policy actions | Verbose prose or uncalibrated "looks good" self-assessments      | Binary pattern match with high false-positive/negative rates on prose |
| Latency & token profile | Purpose-built judgment calls designed for fast, repeatable gating loops             | High token cost and multi-second latency per verification turn   | Sub-millisecond, suitable only for strict syntactic formats           |
| MCP host integration    | Drop-in stdio MCP server on PyPI for Claude Code, Codex, and Cursor                 | Ad-hoc custom prompt templates duplicated across repos           | Custom glue scripts per repository                                    |

Direct answers

## Frequently asked questions

What is Jev Judge MCP in 45 words?

Jev Judge MCP is an open-source Model Context Protocol server published by PyModel LLC on GitHub and PyPI that gives coding and workflow agents independent verify, screen, classify, rerank, compare, review, score, and gate tools backed by TypeSafe AI’s calibrated Jev model.

Why shouldn’t an AI agent grade its own patches or completion claims?

When the same generative model writes a patch and judges whether it met every requirement, confirmation bias and context drift cause it to rubber-stamp partial work. Jev Judge MCP routes verification to an independent typed judgment call that returns calibrated probabilities and policy actions.

How does policy gating (auto, review, escalate) work in Jev Judge MCP?

Instead of parsing free-form prose from a chat prompt, Jev Judge MCP maps calibrated model confidence scores against explicit risk thresholds—allowing high-confidence safe cases to proceed automatically while routing uncertain or high-impact cases to human review or escalation.

Where is Jev Judge MCP published and which hosts support it?

Jev Judge MCP is open source at https://github.com/PyModel/jev-judge-mcp and installable from PyPI (jev-judge-mcp) for Claude Code, Codex, Cursor, and any standard stdio-compatible MCP client.

## Related engineering notes & systems

Architecture write-ups, adjacent products, and delivery services from PyModel LLC.

[Design MCP **Niblet MCP — real product UI references for coding agents** Explore →](/products/niblet-mcp) [Engineering blog **Grounded answers need more than retrieval: building Pixie Q** Explore →](/blog/grounded-answers-pixie-q) [Service pillar **Production AI systems, task evaluations & guardrails** Explore →](/services/production-ai-systems) [macOS utility **ProxyPort for Mac — manage local dev servers started by agents** Explore →](/products/proxyport)

## Scope this workflow with PyModel.

Share your current architecture, data boundaries, and target outcome with our engineering team.

[Start a conversation →](/contact)
