What is Synthezia?
Synthezia is a local-first desktop application for importing audio or video, recording from a microphone, transcribing speech, and working with AI-assisted summaries and chat.
BetaSyntheziaWorkflows, release artifacts, and platform compatibility are still being validated for the public beta.
What you can do
Section titled “What you can do”- Import a local media file or create a microphone recording.
- Let Synthezia normalize the audio and transcribe it with a local Whisper model.
- Review the transcript and add session context such as a title, date, participants, or notes.
- Generate a structured summary, ask questions about that session, or use Global Chat across indexed sessions.
Choose the processing boundary
Section titled “Choose the processing boundary”Local Mode uses local Whisper and Ollama models. After the required models are installed, these processing steps can run without an external AI provider.
External API is opt-in. It routes supported summaries, chat, and embedding operations to the OpenAI-compatible endpoint you configure. That can send transcript, prompt, conversation, and retrieved-context data to that provider.
Important limits
Section titled “Important limits”Synthezia is a beta product. Generated text can be incomplete or wrong and must be reviewed before you use it for a decision or record. The current application does not provide a verified speaker-diarization workflow or general transcript export.