mcpField
G
god-tibo-imagen

god-tibo-imagen

github.com/akillnessCreative & media53260 installsUpdated 4d ago
Repository

About god-tibo-imagen

god-tibo-imagen is a skill and image-generation utility for Codex users, available through a CLI, Node.js library, and Python SDK. It reuses local Codex authentication and supports reference images and dry-run validation.

Key features

  1. 01Generate images from prompts
  2. 02Reuse local Codex authentication
  3. 03Pass multiple reference images
  4. 04Validate requests with dry-run
  5. 05CLI, Node.js, and Python interfaces

Use cases

  1. 01Generating images from shell scripts
  2. 02Creating images with visual references
  3. 03Adding image generation to agent pipelines

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/akillness/jeo-skills --skill god-tibo-imagen

After installing, type/god-tibo-imagenin Claude Code to confirm it loaded.

Bundled filesSKILL.toon

Details

From the project README

god-tibo-imagen — AI Image Generation via Codex Backend

Zero dependencies. Reuses your Codex login. Works from CLI, Node.js, or Python.

god-tibo-imagen generates images by sending requests to Codex's ChatGPT backend. It reads existing local authentication from ~/.codex/auth.json — no separate API key setup required if you already use Codex CLI.

Installation

Plugin (Claude Code)
claude plugin marketplace add NomaDamas/god-tibo-imagen
npm (CLI — global)
npm install -g god-tibo-imagen
# Provides the `gti` command
npm (Library)
npm install god-tibo-imagen
Python SDK
pip install god-tibo-imagen
Skill (any platform)
npx skills add https://github.com/akillness/jeo-skills --skill god-tibo-imagen

Requirements

  • Node.js 20+ (CLI / library)
  • Python 3.10+ (Python SDK)
  • Existing Codex CLI login at ~/.codex/auth.json

When to use

  • Generate AI images without a separate API key — reuses Codex auth
  • Automate image generation from shell scripts, Node.js, or Python workflows
  • Use reference images as context for guided generation (up to multiple images)
  • Run dry-run validation before submitting real generation requests
  • Integrate image generation into agent pipelines via the Python Client class

Do not use when

  • You need a supported public API (this uses an undocumented internal backend — may change)
  • You do not have an active Codex / ChatGPT subscription with a valid ~/.codex/auth.json
  • You need guaranteed SLA or production-grade image generation → use DALL-E API or Stable Diffusion

CLI usage

# Basic generation
gti --prompt "blue square icon on white background" --output ./out.png

# With reference image(s)
gti --prompt "make it round" --input ./ref1.png --input ./ref2.jpg --output ./out.png

# Dry-run (validate without generating)
gti --prompt "sunset over mountains" --dry-run

# Debug output (redacts sensitive data)
gti --prompt "cartoon cat" --output ./cat.png --debug

Node.js library

import { createProvider, resolveConfig } from 'god-tibo-imagen'

const config = await resolveConfig()
const provider = createProvider(config)
const result = await provider.generate({ prompt: 'abstract art' })

Python SDK

from god_tibo_imagen import Client

client = Client()
client.generate_image(
    prompt="blue square icon",
    output="./out.png"
)

Supported formats

FormatInputOutput
PNG✅✅
JPG✅✅
GIF✅—
WebP✅—

Operating rules

  1. Verify ~/.codex/auth.json exists and is valid before running
  2. Use --dry-run to validate prompt and config without consuming quota
  3. Treat the backend as undocumented — avoid hard dependencies in production code
  4. Use --debug to inspect request shape; sensitive tokens are automatically redacted
  5. Pass multiple --input flags for multi-image reference context

References

Source: NomaDamas/god-tibo-imagen — MIT License

Comments

Markdown supported