Featherless

A serverless AI inference platform that provides instant access to tens of thousands of open-source models through a single API, solving the problem of deploying and running open models easily. It targets developers, engineering teams, and operations teams in companies of all sizes, from startups to enterprises. Positioned as a neutral layer independent of any hyperscaler or proprietary ecosystem, it offers a unified API for any open model, with a free tier and dedicated GPU options.

Key features

• Access 40,000+ open-source models
• Single API key for all models
• Serverless inference
• Dedicated GPU options
• Cost calculator
• Model library with categories
• Documentation and quick start guide
• Community Discord support