What Whisper is Whisper is OpenAI's speech-to-text model — it turns audio of spoken language into text. Crucially, it is not a general-purpose conversion model: it does not handle image-to-text, a distinction that mattered in this material because using the wrong model for a task produced a bug.[^01a0f15b-0757-72c4-9201-53d55ca674c7@657] Why the distinction matters When a pipeline needs both kinds of conversion — for example, transcribing audio and reading text out of an image — each step must call the model suited to it. In the captured material, the wrong-model choice (calling Whisper for an image-to-text job) surfaced as a wrong-model bug that had to be debugged during the build.[^01a0f15b-0757-72c4-9201-53d55ca674c7@657] - Whisper is OpenAI's speech-to-text model.[^01a0f15b-0757-72c4-9201-53d55ca674c7@657]- It does not do image-to-text.[^01a0f15b-0757-72c4-9201-53d55ca674c7@657]- Using it for an image-to-text task caused a wrong-model bug in the captured build.[^01a0f15b-0757-72c4-9201-53d55ca674c7@657]