A system is disclosed for facilitating free form dictation, including directed dictation and constrained recognition and/or structured transcription among users having heterogeneous native (legacy) protocols for generating, transcribing, and exchanging recognized and transcribed speech. The system includes at least one system transaction manager having a “system protocol,” to receive a verified, streamed speech information request from at least one authorized user employing a first legacy user protocol. The speech information request which includes spoken text and system commands is generated using a user interface capable of bi-directional communication with the system transaction manager and supporting dictation applications, including prompts to direct user dictation in response to user system protocol commands and systems transaction manager commands. A speech recognition and/or transcription engine (ASR), in communication with the systems transaction manager, receives the speech information request from the system transaction manager, generates a transcribed response, which can include a formatted transcription, and transmits the response to the system transaction manager. The system transaction manager routes the response to one or more of the users employing a second protocol, which may be the same as or different than the first protocol. In another embodiment, the system employs a virtual sound driver for streaming free form dictation to any ASR, regardless of the ASR's ability to recognize and/or transcribe spoken text from any input source such as, for example, a live microphone or line input. In another embodiment, the system employs a buffer to facilitate the system's use of ASRs requiring input data to be in batches, while providing the user with an uninterrupted, seamless dictating experience.