/LLM-models

This repository hosts an automated workflow to generate a curated models.json file. It uses GitHub Actions to periodically fetch the latest Large Language Models from multiple providers (like OpenRouter and XAI), applies advanced filtering, and publishes a clean, structured list for a desktop AI assistant.

Primary LanguagePython

LLM Models List Generator

This repository contains an automated workflow to generate a curated models.json file for the Gemini AI Assistant desktop application.

How It Works

  1. Scheduled Execution: A GitHub Actions workflow runs automatically every 6 hours.
  2. Data Fetching: A Python script (scripts/update_models.py) fetches the latest model data from the OpenRouter API and the official XAI API.
  3. Filtering & Curation: The script applies a series of advanced filters to select only high-quality, multimodal models that are suitable for the assistant's needs. It excludes preview versions, specialized models, and those with insufficient context length.
  4. Structuring: The final list is structured into a clean models.json file, organizing models by provider.
  5. Deployment: The generated models.json is automatically published to GitHub Pages, providing a stable, public URL for the main application to consume.

This approach ensures that the AI assistant always has access to an up-to-date list of the best available models without requiring an application update.