C++ API Documentation Missing / Insufficient for New Users

Where is the C++ library documentation for the advertised library here at the top of the README for this repo?

"CTranslate2 is a C++ and Python library for efficient inference with Transformer models."

You link documentation here for the API(s), then when you visit the link there is only the Python library documentation linked as shown in the included screenshot. Where is the comprehensive C++ library documentation? How are developers supposed to know how to use this library?

For someone new to CTranslate2, “use the examples and read the headers” really isn’t enough to build a real application, especially for more advanced use cases like high-performance speech recognition with Whisper. The examples only cover a few happy-path scenarios and don’t explain the overall design of the library, how different components relate to each other, or recommended patterns for things like batching, streaming, concurrency, memory management, or device/compute configuration. The headers expose a lot of types and options, but without narrative documentation or usage guidance it’s difficult to know which knobs are safe to touch, what defaults are, what is stable vs. internal, or how to compose the pieces correctly for production workloads. As a result, a developer who wants to take advantage of CTranslate2’s performance optimizations in C++ is left guessing about critical behavior, which is risky for correctness and performance. Some dedicated C++ API documentation, even if initially minimal, would go a long way toward making the library approachable and usable beyond the simplest examples.

Start with the README’s C++ library description and the linked CTranslate2 documentation page, then review the existing C++ examples and headers referenced in the issue. Define the documentation scope around API usage, library design, batching, streaming, concurrency, memory management, and device or compute configuration; done means new C++ users have dedicated guidance beyond the Python documentation.