ray.serve.llm.LLMRouter.tokenize

async LLMRouter.tokenize(body: TokenizeCompletionRequest, request: fastapi.Request) → starlette.responses.Response

Tokenize text into token IDs.

This endpoint tokenizes the provided text prompt and returns the token IDs, counts, and optionally token strings.

Note: This is a vLLM specific endpoint.

Parameters:

• body – The tokenize request containing the text to tokenize.

• request – The raw FastAPI request object.

Returns:

A response object with token IDs and metadata.