iMontage RunningHub

AI-powered batch image generation interface using Google Gemini for prompt engineering and RunningHub's FLUX.1-dev LoRA pipeline.

A self-hosted SvelteKit web application that automates high-quality image generation. It uses Google Gemini to craft diverse, scenario-driven prompts from a subject description, then submits them to RunningHub for FLUX.1-dev LoRA-based image synthesis — all through a clean browser UI.

Features

• AI Prompt Engineering — Google Gemini generates unique, detailed FLUX.1-dev prompts from 10 curated photographic scenarios, avoiding repetitive outputs
• Batch Generation — Queue up to 50 sequential image generations with a single click
• LoRA Support — Point to any custom LoRA model URL for character-consistent output
• Flexible Aspect Ratios — 9 presets (1:1, 16:9, 9:16, 3:2, etc.) with automatic 16px-aligned dimension calculation
• Live Status Tracking — Real-time polling with per-image status badges (Processing, Success, Failed, Cancelled)
• Image Preview & Fullscreen — Inline thumbnails with click-to-enlarge modal viewer
• Auto File Management — Sequential filename numbering with configurable prefix and output directory
• Docker-Ready — Multi-stage build with volume-mounted output directory

Architecture

The application runs as a single SvelteKit container with three server-side API routes. The frontend communicates with two external APIs — Google Gemini for prompt generation and RunningHub for image synthesis — while serving generated images from a Docker-mounted volume.

Data Flow

1. User provides subject characteristics, LoRA URL, and API keys
2. A random photographic scenario is selected from 10 curated options
3. Gemini generates a detailed FLUX.1-dev prompt combining the subject and scenario
4. Image dimensions are calculated from the chosen aspect ratio (16px-aligned)
5. The prompt, LoRA, dimensions, and a random seed are submitted to RunningHub
6. The client polls the task status every 5 seconds
7. On success, the image is downloaded and saved to the mounted volume
8. The image and its prompt are displayed in the results feed

Quick Start

Prerequisites

• Node.js 20+
• Docker & Docker Compose (for containerized deployment)
• A Google Gemini API key
• A RunningHub API key

Run with Docker (Recommended)

# Clone the repository
git clone <repo-url>
cd rhub

# Ensure the shared Docker network exists
docker network create shared_net 2>/dev/null || true

# Build and start
docker compose up -d

The interface is available on port 3000 within the shared_net Docker network. Use a reverse proxy (e.g., Nginx Proxy Manager) to expose it externally.

Run Locally (Development)

npm install
npm run dev

Open http://localhost:5173 in your browser.

Configuration

Docker Volumes

API Reference

POST /api/generate

Generates a prompt via Gemini and submits an image generation job to RunningHub.

Request Body:

{
  "loraUrl": "https://...",
  "subject": "Description of the subject",
  "aspectRatio": "1:1",
  "geminiKey": "...",
  "rhubKey": "..."
}

Response:

{
  "taskId": "abc123",
  "prompt": "Generated FLUX.1-dev prompt text..."
}

POST /api/check

Polls RunningHub for task status and downloads the image on success.

Request Body:

{
  "taskId": "abc123",
  "rhubKey": "...",
  "outputDir": "generations",
  "prefix": "image"
}

Response:

{
  "status": "SUCCESS",
  "filename": "image_001.png"
}

Status values: SUCCESS, PROCESSING, FAILED

GET /api/images/[...path]

Serves generated images from the mounted volume. Supports .png and .jpg files with a path traversal security check.

Project Structure

rhub/
├── src/
│   ├── routes/
│   │   ├── +page.svelte              # Main UI (form, results, modal)
│   │   ├── +layout.svelte            # Root layout
│   │   └── api/
│   │       ├── generate/+server.ts   # Gemini prompt gen + RunningHub submit
│   │       ├── check/+server.ts      # Task polling + image download
│   │       └── images/[...path]/+server.ts  # Static image serving
│   ├── app.html                      # HTML shell
│   └── app.d.ts                      # SvelteKit type declarations
├── static/                           # Favicons and web manifest
├── Dockerfile                        # Multi-stage Node.js 20 build
├── docker-compose.yml                # Production deployment config
├── svelte.config.js                  # SvelteKit + adapter-node config
├── vite.config.ts                    # Vite configuration
└── package.json                      # Dependencies and scripts

Tech Stack

• Frontend: Svelte 5 with runes ($state, $props)
• Framework: SvelteKit with adapter-node
• AI: Google Generative AI SDK (@google/generative-ai) — Gemini 3 Pro
• Image Gen: RunningHub API — FLUX.1-dev with LoRA
• Runtime: Node.js 20 (multi-stage Docker build)
• Build: Vite 6, TypeScript 5

Scripts

npm run dev        # Start dev server (hot reload)
npm run build      # Production build
npm run start      # Run production build
npm run preview    # Preview production build locally
npm run check      # TypeScript and Svelte type checking

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/my-feature)
3. Commit your changes (git commit -m 'Add my feature')
4. Push to the branch (git push origin feature/my-feature)
5. Open a Pull Request

License

Private — All rights reserved.