Zilla

Zilla is an open-source large language model (LLM) gateway that accepts API requests for OpenAI, Anthropic, Gemini, and AWS Bedrock models, routing and translating between different provider APIs. It features streaming support, tool calls, guardrails for prompt filtering, model fallback mechanisms, and token usage metrics. It enables developers to unify access to multiple LLM providers through a single interface without storing user API keys.