Doubao Say v1.3.1
Dictate into Linux applications using configurable tap or hold shortcuts.
Before you upgrade to v1.3.1
v1.2.0 requests thinking off where supported; `glm-5.3` and `glm-5.3-flash` instead request `reasoning_effort: low`. The five-second fallback remains.
What changed in v1.3.1
The v1.0.0–v1.3.1 launch sequence adds native X11 paste, Deepgram US English recognition and configurable Vibekey controls. Later releases refresh microphone devices, finish Fn dictation with another key and restore plain text ahead of HTML.
Also: Native X11 paste, optional Volcengine and Deepgram recognition, experimental polishing, Vibekey controls, device recovery and clipboard restoration.
v1.3.1 app or Omarchy-plugin archives support Python 3.11–3.14; choose one route. GTK, WebKit and PipeWire are external dependencies. Git-installed plugins must be disabled before updating, checked, then re-enabled.
New things you can do
Available at launch; v1.1.0 buffers initial audio locally and discards unconfirmed gestures without uploading.
Read the evidence
Live partial text is shown during recording and the final transcript arrives after recording stops.
Read the evidence
Added in v1.3.0; the API key is stored separately with owner-only permissions.
Read the evidence
Clipboard paste is the default; unsupported desktops and unknown targets keep results for manual copying.
Read the evidence
Added in v1.1.0 with terminal-aware paste shortcuts and an overlay that does not request activation.
Read the evidence
Optional direct typing sends text gradually and stops remaining input on cancellation or detected focus loss.
Read the evidence
Presets or a custom chord apply immediately; hold and double-tap timings are configurable.
Read the evidence
Pause previews do not end recording or paste; v1.2.0 adds a separate status row and provider-aware reasoning guidance.
Read the evidence
v1.2.0 rescans inputs when Settings opens or devices are refreshed, preserving a temporarily unavailable saved device.
Read the evidence
The tray exposes the recent result for copying; recognition timeouts retain partial text without automatic paste.
Read the evidence
v1.1.0 adds conditional restoration; v1.3.1 prefers plain text over HTML to prevent markup in subsequent pastes. Images, file lists and HTML-only payloads retain their handling.
Read the evidence
All six controls are configurable; defaults cover recording, Enter, cancel, Up, Down and Meta+Backspace.
Read the evidence
Four waveform styles and an appearance preview are available.
Read the evidence
Language choices save and apply automatically.
Read the evidence
A red overlay indicator and control-center update button identify a newer release.
Read the evidence
Fixed and improved
v1.1.0 refreshes setup readiness after checks; v1.2.0 stops recording when the microphone disconnects, and a failed keyboard no longer drops all keyboard listeners.
Seeded by @iluxav with Codex, from the release's public sources. Each change links to its evidence.