What it is
EKHOS AI is a desktop transcription application that converts audio and video into text using locally run artificial intelligence models. The product is presented as a professional-grade transcription assistant aimed at users who require accuracy and local data handling, including legal teams, medical professionals, law enforcement, journalists, researchers and educators.
The core capability is offline, on-device transcription that supports real-time microphone capture and transcription of audio played by the computer. It includes automatic speaker recognition and labeling, support for 98 languages, and the ability to process many common media formats such as MP3, MP4, WAV, MKV, AVI and MPEG.
The system runs its AI models on the user’s machine and can use either CPU or GPU resources (NVIDIA GPUs recommended). Users can select between multiple AI processing models and transcription modes to balance speed and accuracy.
Key features
Offline local processing and unlimited transcription are central features: files are processed and stored on the local machine with no external transmission asserted. The application supports bulk queueing of files for batch transcription and handles large or long media without imposed size or duration limits.
Audio capture options include direct microphone input and speaker capture for transcribing meeting audio or media played on the computer. Speaker identification and labeling are provided to separate and tag voices within transcripts.
Security and access controls include local password protection, a recovery phrase mechanism, transcript encryption in a local database, masking and read-only transcript protection, and claims that no telemetry or usage analytics are sent externally and that transcripts are not used to train external models.
Proofreading and export tools include a built-in media player, a Tracks Editor that highlights transcript segments in sync with playback, a search facility, a plain text editor alternative, and export to Word, PDF and plain text formats with clipboard copy support. Personalized settings and transcript management (update, save, export, delete, search) are also provided.
Use cases
Legal professionals can use EKHOS AI to transcribe depositions, interviews and courtroom recordings while retaining local control of sensitive audio and text files. Speaker labeling supports assigning statements to different participants for case review.
Medical practitioners can transcribe consultations and dictated notes for record-keeping or clinical documentation, using local processing to limit external data exposure. Researchers and journalists can transcribe interviews, focus groups and field recordings, using bulk queueing and export formats for downstream analysis.
Educators, podcasters and meeting hosts can convert lectures, episodes and meeting audio into searchable transcripts, make timestamped edits via the Tracks Editor, and export transcripts for distribution or archiving. GPU acceleration is available where faster processing is needed.
