Multi-modal input on an electronic device

A system comprising:

• a data processing apparatus;
• and
• a data store coupled to the data processing apparatus, in which is stored:
• an application-independent input method editor configured to receive input for a plurality of applications executable by an electronic device, the application-independent input method editor operable to:
• receive spoken input from a user of the electronic device that is directed to an application of the plurality of applications;
• generating, with the electronic device, a correlation between the spoken input and the application, from the plurality of applications, to which the spoken input is determined to correspond, wherein generating the correlation comprises (i) determining an application type for the application that identifies a category of application to which the application is assigned, each category of application capable of including a plurality of different applications, and (ii) matching one or more utterances in the spoken input to one or more stored terms that are associated with the application type;
• provide the spoken input and the generated correlation to a server, wherein the server is remote to the electronic device and includes a speech recognition system configured to access one or more language models for recognizing text based on the spoken input and the correlation;
• receive text from the server, wherein the text represents a translation of the spoken input;
• select an application, from the plurality of applications, to receive the text, the selection using the generated correlation;
• and
• provide the text as input to the selected application.

A non-transitory computer-readable storage device encoded with a computer program product, the computer program product including instructions that, when executed, cause data processing apparatus to perform operations comprising:

• receiving spoken input from a user of an electronic device in an application of the plurality of applications;
• generating, with the electronic device, a correlation between the spoken input and the application, from the plurality of applications, to which the spoken input is determined to correspond, wherein generating the correlation comprises (i) determining an application type for the application that identifies a category of application to which the application is assigned, each category of application capable of including a plurality of different applications, and (ii) matching one or more utterances in the spoken input to one or more stored terms that are associated with the application type;
• providing the spoken input and the generated correlation to a server, wherein the server is remote to the electronic device and includes a speech recognition system configured to access one or more language models for recognizing text based on the spoken input and the correlation;
• receiving text from the server, wherein the text represents a translation of the spoken input;
• selecting an application, from the plurality of applications, to receive the text, the selection using the generated correlation;
• and
• providing the text as input to the application.