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

codex-gemini-jev-bridge

Local MCP bridge letting Codex hand drafting to Gemini via a Google AI Pro quota, with Jev checking drafts against explicit criteria.

MCP serverbetaMIT
View on GitHub ↗

Overview

Gemini + Jev Bridge for Codex is a local MCP server that lets Codex delegate substantial routine drafting, analysis and code proposals to Gemini. Codex plans the work, reviews and applies changes, runs tests and gives the final answer; Jev checks each Gemini draft against explicit criteria.

Gemini runs on the included quota of a Google AI Pro subscription through Antigravity CLI, with no API key, billing, overages or paid fallback; when quota runs out, Gemini work stops until reset. Jev uses a separate TypeSafe account. Tools include delegate_task (Gemini drafts text or a patch, Jev checks one to six criteria, results saved locally) and jev_evaluate (up to 12 compact routing, ranking or evidence questions per request). Only task text and explicitly selected files are sent to providers. Guides are in English and Russian, including a setup prompt for another GPT or Codex chat.

What the decision model decides

TypeQuestionAnswers
Yes / noDoes the Gemini draft meet each explicit criterion?yes · no
ChoiceCompact routing, ranking or evidence questions (jev_evaluate)—

How it works

  1. Codex calls delegate_task
  2. Gemini drafts via Antigravity CLI
  3. Jev checks 1-6 criteria
  4. Codex reviews, applies and tests

Strengths

  • No Gemini API billing
  • Explicit criteria checks on drafts
  • Only selected files are sent

Caveats

  • Depends on Google AI Pro quota terms
  • Instructions written for Windows/PowerShell

Getting started

Follow docs/USER_GUIDE.md: Node.js 22+, Codex with local MCP, Antigravity CLI signed in, and a TypeSafe key.

At a glance

Type
MCP server
Status
beta
Runtime
local MCP server (Node.js 22+)
Has code
yes
License
MIT

Use cases and integrations

How curated lists describe it

  • Local MCP bridge for Codex: Gemini uses included Google AI Pro quota via Antigravity CLI (no API billing or credit overages); TypeSafe Jev reviews drafts. EN/RU user and GPT guides. MIT.

    hellogumbo/awesome-jev

Frequently asked questions

Does Gemini use a paid API?

No, it uses Google AI Pro subscription quota via Antigravity CLI.

What does Jev check?

Gemini drafts against one to six explicit criteria.

Is Jev included in Google AI Pro?

No, it needs a separate TypeSafe key.

Related projects

  • jev-architect (BenjaminPolge)

    Skill that makes Claude Code and Codex split AI features into Jev, LLM or plain-code steps before writing code.

    Projectbeta
  • jevmory

    Local memory for coding agents that stores verbatim quotes and has Jev grade memory lines as keep, stale, wrong or unsupported.

    ProjectbetaPython
  • jevskill

    Agent skill that has Jev pick the few relevant lines from big logs before your coding agent reads them, with a reversible reduce and ledger.

    ProjectbetaPython
  • typesafe-as-a-judge

    Unofficial Codex and Claude Code MCP plugin giving agents bounded Jev tools for routing, ranking, extraction, verification and judging.

    MCP serverbeta
  • jev-agent-skill

    Chinese-language agent skill that offloads classify, screen, score and compliance checks to Jev via OpenCode Zen's free tier.

    PluginbetaPython
  • jev-browser-wingman

    MCP wingman that hands an agent's browser clicks and typing to Jev on a real logged-in Chromium, alongside Playwright MCP.

    MCP serverbetaTypeScript