OpenScreech¶
OpenScreech is a local-first voice work parrot for macOS: built for dictation, transcription, reusable processing pipelines, spoken responses, automation, recordings, and searchable Notes.
Use OpenScreech Home as both a quick-action dashboard and a workspace for its primary tools. Features can stay embedded beside Home's sidebar or move into separate windows without losing their current state. Keep common controls close in the menu bar, or start profile-driven dictation directly from a hotkey. Speech recognition runs locally by default, while optional LLM post-processors can use local or remote endpoints that you choose.
Start Here¶
New to OpenScreech? Follow these guides in order:
- Meet OpenScreech and learn its basic surfaces.
- Complete the first-launch and preflight checks.
- Make your first dictation.
- Continue with After Your First Dictation.
How Dictation Fits Together¶
Hotkey → Profile → Capture and ASR → Processing → Speech, paste, routing, files, or Notes
The hotkey starts and stops dictation and selects the profile for that session. The profile decides how the recognized text is cleaned up, processed, delivered, and stored.
Choose What You Want to Do¶
Dictate into other apps¶
- Configure hotkeys and dictation behavior.
- Follow Dictate into Documents for a practical starting profile.
- Use Dictate Prompts for AI Tools.
- Recall recent results from the overlay with Left and Right Arrow without transcribing or processing them again.
Clean up and organize transcriptions¶
- Apply deterministic cleanup with word replacement libraries.
- Rewrite or format text with LLM post-processing.
- Build sequential or branched multi-stage LLM workflows.
- Capture and Organise Voice Notes to keep ideas searchable by text, folder, and tag.
Automate voice workflows¶
- Run matched actions with trigger routing libraries.
- Run Tools with Voice Commands using spoken trigger phrases.
- Add Transcription to Your App using the local HTTP service.
Hear a response¶
- Turn a final result or individual LLM stage into local speech with Screechback.
- Follow the Hear Summaries and Responses preset.
- Download and test speech models and complete voice catalogues before enabling a profile.
Work with recordings and files¶
- Record a Podcast with separate microphone tracks and individual or combined transcripts.
- Record and Summarise Meetings to retain the discussion and review decisions and action items.
- Transcribe Audio and Video Files, from one existing recording to a whole folder.
Find Settings and Files¶
- Use the menu and window map to locate controls.
- Look up unfamiliar language in the glossary.
- Check data and output locations to understand what OpenScreech stores and where.
- Review output formats when choosing transcript artifacts.
- Start with troubleshooting when something is not working as expected.
OpenScreech keeps speech recognition local by default and exposes advanced behavior explicitly through profiles, reusable libraries, and workflow-specific settings.
Quick Tips¶
- OpenScreech Home does not need to remain open for dictation. Once configured, a hotkey can start the complete capture and paste workflow from another application.
- When the overlay says
Listening (Profile Name), it is confirming which profile will handle the result if you finish at that moment.