via “unified-llm-api-abstraction-with-provider-detection”
Python SDK, Proxy Server (AI Gateway) to call 100+ LLM APIs in OpenAI (or native) format, with cost tracking, guardrails, loadbalancing and logging. [Bedrock, Azure, OpenAI, VertexAI, Cohere, Anthropic, Sagemaker, HuggingFace, VLLM, NVIDIA NIM]
Unique: Implements provider detection via regex-based model name matching and a centralized provider configuration registry that maps 100+ models to their native APIs, with automatic request/response translation using provider-specific handler classes rather than a single generic adapter
vs others: More comprehensive provider coverage (100+ vs ~20-30 for competitors) and automatic provider detection without explicit configuration, reducing boilerplate compared to LangChain or raw SDK usage