Note to developers regarding Apple’s Speech Recognition Toolbox:
Apple’s proprietary Speech Recognition Toolbox has been private until now. The only shipping program that has had access to the API so far is the user interface known as Casper, which uses the Speech Recognition Toolbox to recognize menu items, dialog buttons, and AppleScript-like macro names, and then uses AppleScript and Text-to-Speech Toolboxes to map the recognition results into appropriate actions and feedback.
By using the Speech Recognition Toolbox directly through it’s API, developers can add speech recognition features to their applications relatively easily, while using less memory and benefiting from more efficiency and accuracy than possible using Speech Recognition indirectly through the Casper user interface and AppleScript.
The preliminary documentation accompanying this note describes the Speech Recognition Toolbox, including what it can do, its limits, and how applications can use it. (The document is a MacWrite Pro document.) Parts of the Speech Recognition Toolbox are still in flux. (For example, we anticipate adding support for a standard user feedback module, and standard use of keywords and rejection levels.) Furthermore, this documentation is very preliminary. We anticipate improving it greatly based on feedback from the few selected developers who will get an early look at the Speech Recognition Toolbox API. We also anticipate working closely with a few early developers to help them figure out the best ways to use the toolbox in their applications.
Apple’s Speech Recognition Toolbox and API are not public yet. Please do not make copies or distribute any of this documentation. You can AppleLink PLAINTALK if you have questions or comments.
