feat(local): local speech recognition, for recordings as sources and voice input by MODSetter · Pull Request #2264 · MODSetter/SurfSense

feat(models): a Speech recognition section in Settings, and the set-up popover

MODSetter

committed

b8e3c5d

feat(local): stage speech recognition's audio.cpp server, not yet started

MODSetter

committed

5d4dc25

fix(models): a speech section that fails honestly and never waits on servers

MODSetter

committed

4397a53

fix(local): never leave a second speech recognition server running

MODSetter

committed

75603e3

feat(sources): recordings in the sources panel, with how far their transcript is

MODSetter

committed

e6fbf2b

fix(sources): keep a recording's set-up open, and say when its language is unlisted

MODSetter

committed

057fa19

feat(microphone): grant the mic to the app's own page, and bridge the system's switch

MODSetter

committed

f8e5ba0

feat(local): offer Nemotron 3.5 ASR for a new speech recognition slot

MODSetter

committed

a08ef99

fix(microphone): let CI run the macOS check, and note the sidecars' entitlement

MODSetter

committed

b193fcf

fix(local): refuse a listed speech language with no locale for its family

MODSetter

committed

88fd9ad

feat(dictation): voice input in the composer, a phrase at a time

MODSetter

committed

8b3dd42

fix(dictation): never send the quiet alone, and say when memory stopped a phrase

MODSetter

committed

f144804

feat(local): transcribe speech through one gate in the API

MODSetter

committed

939de65

fix(local): answer a hung speech server with 503 and keep numbers in CJK text

MODSetter

committed

6c952c2

feat(local): transcribe recordings as sources in the ingest worker

MODSetter

committed

553048c

fix(local): decode Ogg recordings under any path and clear a stopped recording's audio

MODSetter

committed

0de99e8

Merge branch 'feat/local-speech-recognition-ui' into feat/local-speech-recognition

MODSetter

committed

5f393ac

Merge remote-tracking branch 'origin/dev' into feat/local-speech-recognition

MODSetter

committed

d9c0a86

fix(local): keep a ChatGPT plan out of speech settings, and docs that match v1

MODSetter

committed

531c1f6

Navigation Menu

Repository navigation

feat(local): local speech recognition, for recordings as sources and voice input - #2264

feat(local): local speech recognition, for recordings as sources and voice input#2264
MODSetter merged 19 commits into
devfrom
feat/local-speech-recognition

Commits on Oct 14, 2026

feat(models): a Speech recognition section in Settings, and the set-up popover

feat(local): stage speech recognition's audio.cpp server, not yet started

fix(models): a speech section that fails honestly and never waits on servers

fix(local): never leave a second speech recognition server running

feat(sources): recordings in the sources panel, with how far their transcript is

fix(sources): keep a recording's set-up open, and say when its language is unlisted

feat(microphone): grant the mic to the app's own page, and bridge the system's switch

feat(local): offer Nemotron 3.5 ASR for a new speech recognition slot

fix(microphone): let CI run the macOS check, and note the sidecars' entitlement

fix(local): refuse a listed speech language with no locale for its family

feat(dictation): voice input in the composer, a phrase at a time

fix(dictation): never send the quiet alone, and say when memory stopped a phrase

feat(local): transcribe speech through one gate in the API

fix(local): answer a hung speech server with 503 and keep numbers in CJK text

feat(local): transcribe recordings as sources in the ingest worker

fix(local): decode Ogg recordings under any path and clear a stopped recording's audio

Merge branch 'feat/local-speech-recognition-ui' into feat/local-speech-recognition

Merge remote-tracking branch 'origin/dev' into feat/local-speech-recognition

fix(local): keep a ChatGPT plan out of speech settings, and docs that match v1