Repository navigation
Add Hey Nativ wake-word dictation - #567
Merged
Merged
Conversation
Confirm HN-2 candidates through the selected speech-to-text pipeline while retaining contiguous microphone audio. Add bounded pre-roll, confirmation reuse, adaptive ambient-noise gating, model fixtures, and capture regression coverage.
Blaizzy
changed the title
Blaizzy
changed the title
Blaizzy
marked this pull request as ready for review
September 21, 2026 22:25
Blaizzy
force-pushed
the
pc/hey-nativ-wake-word
branch
from
September 21, 2026 22:37
08c8813 to
437d0ea
Compare
lucasnewman
approved these changes
Sep 21, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds opt-in Hey Nativ dictation: say the wake phrase and continue speaking. Two seconds of silence or the shortcut finishes capture; the wake phrase is removed before insertion.
An adaptive energy gate activates the bundled HN-2 classifier on CPU/Neural Engine. The selected local speech model confirms candidates while audio continues buffering. Requires a running Nativ server and an installed speech model.
Validation
SystemMonitorIdentityTests.swiftbuild issue.Battery measurements
12 unplugged trials over 17.8 minutes on an M5 Max; built-in microphone for live trials, optimized native probes, ASR model kept loaded. App UI was excluded.
¹ Percentage of one CPU core. Steady-noise classifier calls fell from approximately 50/sec to zero after warm-up. Battery charge displayed 100% → 99% across the mixed-condition session.
Power includes desktop activity; live audio varied between trials. Battery telemetry refreshed about once per minute, providing 2–4 fresh readings per condition.