Welcome
Boundless Flow is a desktop speech workstation with a focus on Chinese-language use.
Screenshots show the app interface with example settings and sessions. Use your own models, services and recordings.
First time here? Start with a taskFollow illustrated tutorials from installation and model setup to your first result →Four everyday views
| Feature / message | Guidance |
|---|---|
| Live | Read the transcript and translations. Pause following to browse earlier text, or open Focus Reading. |
| Meetings & notes | Open saved meetings, find passages by time or speaker, edit the transcript and create minutes. |
| Voiceprint | Add reference recordings, compare candidates and check calibration. Review speaker labels separately from identity matches. |
| Raw | Review the transcript, proofread it, then copy or export. |
- 1Start with the setup guide.
- 2Start recording after setup.
- 3Open saved meeting records here.
Tools and operating modes
Use the header to open the setup wizard, recording settings, voice tools, interpretation or floating transcription. The voice panel has six task entries for audio files, model setup, timeline speaker names, speech synthesis, audio output and emotion analysis. Choose the one you need.
Local models need complete packages, matching native components and enough memory. Model services need a running server and the matching protocol. Cloud APIs need an account with the appropriate capability. Start with one route; there is no need to configure every service.
- 1Choose the family matching your downloaded model.
- 2Select the complete model directory.
- 3Choose a clear single-speaker reference.
Before you start
This guide covers the 0.5.0 desktop app. Check model files and service connections, then test recording, translation and speech on your devices. Follow the on-screen guidance for experimental or unavailable features. Recognition quality and interpretation latency vary with models, languages and devices.