SpeechPulse

SpeechPulse is a desktop application that provides speech recognition and translation capabilities for dictation and transcription. It targets individuals who need to turn spoken language into text across their preferred programs and environments, including writers, professionals, and privacy-conscious users. The software emphasizes local processing and wide language support as part of its core functions.

SpeechPulse operates by transcribing spoken input into text in real time, with options for translating speech and handling punctuation. It supports multiple languages and can work with both mic input and system audio in various modes. The tool is designed to integrate with many applications and input areas, aiming to streamline voice-to-text workflows across different contexts.

The product includes features such as automatic and manual punctuation, push-to-talk input, and AI templates that work with language models and APIs for editing and formatting. It also offers a mode that processes audio files with speaker diarization and can generate subtitles. The software can download additional speech models and GPU libraries to improve performance on compatible systems.

SpeechPulse provides information about updates, installation, and model management through its help and update sections. It notes compatibility considerations, such as requiring NVIDIA GPUs for certain live-dictation models and the need to handle CUDA libraries.

Use cases

• Dictation and transcription across apps
• Audio and video subtitle generation