# LLMCapabilities Capabilities and features supported by an LLM service - `supportsChat` (`boolean | null`, optional): Supports conversational/chat completion format with message roles - `supportsCompletion` (`boolean | null`, optional): Supports raw text completion with prompt continuation - `supportsFunctionCalling` (`boolean | null`, optional): Supports function/tool calling with structured responses - `supportsSystemMessages` (`boolean | null`, optional): Supports system prompts to define model behavior and context - `supportsStreaming` (`boolean | null`, optional): Supports real-time token streaming during generation - `supportsSamplingParameters` (`boolean | null`, optional): Supports sampling parameters like temperature, top_p, and top_k for generation control [TypeScript](../../typescript.md)