feat: identify speakers in the scene stream (voice-detect.cpp, C-API v9) - #78

Merged

mudler merged 19 commits into

masterfrom

feat/voice-identification

Sep 30, 2026

Commits on Sep 30, 2026

feat(speaker): add SpeakerRegistry for enrolled voices
mudler
committed
e4261a6
fix(speaker): reject a registry blob with dim 0 and speakers
mudler
committed
85af742
feat(speaker): add SpeakerIdentifier for diarization slots
mudler
committed
e22e5e5
fix(speaker): test overlap against closed segments and document the open contract
mudler
committed
f0bd407
feat(speaker): carry speaker names through words, utterances and scene JSON
mudler
committed
5837b4a
feat(speaker): fold voice-detect.cpp in behind SpeakerEncoder
mudler
committed
b6e42cf
feat(speaker): identify speakers inside the scene stream
mudler
committed
b1a9544
fix(speaker): check speaker preconditions first and test order-independent naming
mudler
committed
07a9984
feat(capi): speaker identification, ABI v9
mudler
committed
5c7ef6c
fix(capi): read scene options only within the caller's size and avoid a stream leak
mudler
committed
1f639c5
feat(cli): enroll speakers and name them in scene output
mudler
committed
b9bafb4
fix(cli): write the speaker registry atomically and keep the docs consistent
mudler
committed
7260f7e
fix(speaker): name a slot while its segment is still open
mudler
committed
88500cb
fix(scene): keep one JSON shape for a stream with a speaker part
mudler
committed
c0d5896
fix(speaker): save registries atomically and do not replace an unreadable one
mudler
committed
ca4b586
docs(speaker): give a starting threshold per encoder and gate the voice-detect OFF build
mudler
committed
9b2b5a2
test(speaker): word a comment without an arrow
mudler
committed
29053a9
build: point the voice-detect submodule at localai-org
mudler
committed
4c742ac
build: pin voice-detect.cpp to its merge commit
mudler
committed
4cff7d5

Navigation Menu

feat: identify speakers in the scene stream (voice-detect.cpp, C-API v9) - #78

feat: identify speakers in the scene stream (voice-detect.cpp, C-API v9)#78
mudler merged 19 commits into
masterfrom
feat/voice-identification

Commits on Sep 30, 2026

feat(speaker): add SpeakerRegistry for enrolled voices

fix(speaker): reject a registry blob with dim 0 and speakers

feat(speaker): add SpeakerIdentifier for diarization slots

fix(speaker): test overlap against closed segments and document the open contract

feat(speaker): carry speaker names through words, utterances and scene JSON

feat(speaker): fold voice-detect.cpp in behind SpeakerEncoder

feat(speaker): identify speakers inside the scene stream

fix(speaker): check speaker preconditions first and test order-independent naming

feat(capi): speaker identification, ABI v9

fix(capi): read scene options only within the caller's size and avoid a stream leak

feat(cli): enroll speakers and name them in scene output

fix(cli): write the speaker registry atomically and keep the docs consistent

fix(speaker): name a slot while its segment is still open

fix(scene): keep one JSON shape for a stream with a speaker part

fix(speaker): save registries atomically and do not replace an unreadable one

docs(speaker): give a starting threshold per encoder and gate the voice-detect OFF build

test(speaker): word a comment without an arrow

build: point the voice-detect submodule at localai-org

build: pin voice-detect.cpp to its merge commit