Module count_tokens 

Token-count request/response mappings. Counts always describe the selected target model’s tokenizer; conversion cannot estimate a different tokenizer.

Re-exports

pub use request::*;

Modules

request

Direct counting requests reuse the same pair’s content/tool field mappings. No generation request or invented output token budget is constructed.

Enums

ClaudeCountContext

Facts from the original Claude counting request and any context-edit pass.

Functions

claude_to_gemini_response

claude_to_openai_response

gemini_to_claude_response

gemini_to_openai_response

openai_to_claude_response

openai_to_gemini_response