Image to SVG - Batch Processing

Convert images to SVG using AI. Two-stage pipeline: Gemini analyzes images, GPT-5 generates SVG code. 50% cost savings using AI batch APIs.

Quick Start

# Install

npm install -g llmsvg

# Initialize

llmsvg init

# Configure (update .env file using .env.example, set prompts, etc..)

# create folder <batchname> in input then paste in all input files
mkdir input/<batchname>


# Starts pipeline on <batchname> read all input files
# Automatically checks every 5 minutes for updates from API and proceeds to next stage when all descriptions are ready
# Once complete, descriptions and SVGs will be found in output/<batchname>

llmsvg <batchname>

# If terminal times out, run with -r flag to resume existing job

llmsvg <batchname> -r

# For more help, run llmsvg help

All Commands

llmsvg init              # Set up folders

llmsvg <batch-name>      # Process batch

llmsvg <batch-name> -r   # Resume interrupted job

llmsvg help              # Show help

Folder Structure

project/

├── input/

│ └── my-batch/ # Your images

├── jobs/

│   └── my-batch/ # Job tracking, used to resume job

├── output/

│   └── my-batch/

│       ├── descriptions/ # Gemini output

│       └── SVGs/ # GPT-5 output

└── .env # API keys and config

Made by Hammaad Memon @ RAIVN