
Welcome to ViscribeAI
ViscribeAI is an open-source library for turning images into structured output. Define the result shape, pass an image, choose an AI model, and get parsed data back instead of loose free-form text.Start with the quickstart
Install the Python or TypeScript package and run your first image workflow.
What you can build
Describe images
Generate objective descriptions and optional tags for images.
Extract structured data
Convert receipts, screenshots, forms, product photos, and documents into schema-shaped
output.
Classify images
Route images into allowed labels or free-form categories.
Ask visual questions
Ask focused questions when you need a specific answer from an image.
Compare images
Summarize similarities, differences, variants, and visual changes.
Bring your model setup
Use OpenAI-compatible model configuration, including model names, keys, base URLs,
retries, timeouts, and generation options.
Packages
model_config in Python or modelConfig in TypeScript.
First example
Load provider credentials from your environment when possible. For the default OpenAI
client, set
OPENAI_API_KEY; for compatible providers, pass the relevant client
settings through model_config or modelConfig.Next steps
Quickstart
Install ViscribeAI, configure credentials, and run one image method.
Model Providers
Configure OpenAI-compatible providers, credentials, and request options.
Community
Find ViscribeAI on GitHub, Discord, X, and LinkedIn.
Contributing
Set up the repository and run the checks used by contributors.