Stop writing repetitive custom boilerplate every time a new open-source or proprietary LLM drops. Our lightweight core SDK has been refactored to abstract provider-specific configurations into a single standard payload.

Our lightweight core SDK has been entirely refactored from the ground up to fully abstract provider-specific configurations, distinct API schemas, and tokenization formats into a single, highly optimized standard payload. This allow you to benchmark, test, and switch between OpenAI, Anthropic, DeepSeek, or Llama models instantly.