feat: text to speech (#546) #710

IgorSwat · 2026-01-09T08:46:21Z

Description

Introduces a breaking change?

Yes
No

Type of change

Bug fix (change which fixes an issue)
New feature (change which adds functionality)
Documentation update (improves or adds clarity to existing documentation)
Other (chores, tests, code style improvements etc.)

Tested on

iOS
Android

Testing instructions

Screenshots

Related issues

Checklist

I have performed a self-review of my code
I have commented my code, particularly in hard-to-understand areas
I have updated the documentation accordingly
My changes generate no new warnings

Additional notes

There are still a few side things to be done on this feature.

…-mansion/react-native-executorch into @is/text-to-speech

…ion/react-native-executorch into @is/text-to-speech

Removed unnecessary Log.h include from RnExecutorchInstaller.h

Removed unnecessary iostream include from BaseModel.cpp

msluszniak

Please apply changes to all applicable places, I didn't repeated these comments. But overall, it looks really cool, these suggestions are just very small nits ;)

packages/react-native-executorch/common/rnexecutorch/data_processing/Sequential.h

packages/react-native-executorch/common/rnexecutorch/models/text_to_speech/kokoro/Decoder.cpp

packages/react-native-executorch/common/rnexecutorch/models/text_to_speech/kokoro/Decoder.h

packages/react-native-executorch/common/rnexecutorch/models/text_to_speech/kokoro/Constants.h

...act-native-executorch/common/rnexecutorch/models/text_to_speech/kokoro/DurationPredictor.cpp

...ges/react-native-executorch/common/rnexecutorch/models/text_to_speech/kokoro/Partitioner.cpp

packages/react-native-executorch/common/rnexecutorch/models/text_to_speech/kokoro/Utils.cpp

…ion/react-native-executorch into @is/text-to-speech

msluszniak · 2026-01-09T14:10:03Z

Since you added demo app in this PR, you should also update README.md in the following section:

chmjkb

Left some nits for the moment, I feel like we need to add docs to have a solid understanding of the public API layer. However, this looks like really a lot of solid work, congrats!

Haven't finished reviewing tho :D

chmjkb · 2026-01-09T14:30:32Z

apps/speech/screens/TextToSpeechScreen.tsx

+  const [readyToGenerate, setReadyToGenerate] = useState(false);
+
+  const audioContextRef = useRef<AudioContext | null>(null);
+  const sourceRef = useRef<any>(null);


Can we type this?

chmjkb · 2026-01-09T14:30:49Z

apps/speech/screens/TextToSpeechScreen.tsx

+      iosOptions: ['defaultToSpeaker'],
+    });
+
+    // Initialize context once


Suggested change

// Initialize context once

chmjkb · 2026-01-09T14:33:29Z

apps/speech/screens/TextToSpeechScreen.tsx

+
+      const onEnd = async () => {
+        setIsPlaying(false);
+        setReadyToGenerate(true);


doesnt the useEffect above handle this?

chmjkb · 2026-01-09T14:42:00Z

packages/react-native-executorch/scripts/create-package.sh

Why is this deleted?

chmjkb · 2026-01-09T14:45:44Z

packages/react-native-executorch/src/types/tts.ts

+export enum TextToSpeechLanguage {
+  EN_US = 0,
+  EN_GB = 1,
+}


S2T shared this as string literal union and I recommend using it instead, so we have a unified approach:

export type SpeechToTextLanguage = | 'af' | 'sq' | 'ar' // this continues

chmjkb · 2026-01-09T14:48:00Z

packages/react-native-executorch/src/types/tts.ts

+// Voice configuration
+// So far in Kokoro, each voice is directly associated with a language.
+// The 'data' field corresponds to (usually) binary file with voice tensor.


Can we use JSDoc for this kind of comments?
For example:

** * Voice configuration * * So far in Kokoro, each voice is directly associated with a language. * The 'data' field corresponds to (usually) binary file with voice tensor. */

Overall I think that this is a good approach for everything we share to the public API as it is easy for the user to figure out what the thing is doing without the docs.

chmjkb · 2026-01-09T14:50:18Z

packages/react-native-executorch/src/constants/modelUrls.ts

+export const URL_PREFIX =
  'https://huggingface.co/software-mansion/react-native-executorch';
-const VERSION_TAG = 'resolve/v0.6.0';
-// const NEXT_VERSION_TAG = 'resolve/v0.7.0';
+export const VERSION_TAG = 'resolve/v0.6.0';
+export const NEXT_VERSION_TAG = 'resolve/v0.7.0';


Does this need to be exported? This makes it possible for the users to import this, which is unnecessary

IgorSwat and others added 30 commits December 3, 2025 18:22

implement Kokoro components

fa489ee

restructurize kokoro submodules

2320f72

deps: bump Expo to 54, specify Expo peer deps verisons

5cf5c62

fix: revert legacy imports

a266983

main model logic

716445b

deps: fix example app deps

4bffd23

fix: remove ios cache

26cc16f

fix: bring back expo prebuild cache

3f29191

chore: remove create-package.sh

bf50496

more progress...

9a57da2

kokoro main inference implemented

b2bff9d

text to speech prototype

dd5a001

Merge branch '@chmjkb/expo-54-upgrade' of https://github.com/software…

e3dae81

…-mansion/react-native-executorch into @is/text-to-speech

various fixes & improvements

3cfefd5

Temporary testing screen

c3bee1c

reorganize DurationPredictor data flow

99ffe94

text-to-speech mvp

4173cee

add ios support

1e37875

implement fallback phonemization (US)

b67bc33

fix 'ed' phonemization bug

9aa94ea

add british support

6d11367

add cropping audio vector

6c94783

small refactor

59dc3f6

demo app finished

67ab821

update model input variants

07a2ed2

implement input partitioning

9bea9ba

native side streaming

3308669

audio streaming fixed

5ac719c

enable additional model options

3de1722

fix reload bug

8c171b1

IgorSwat added 18 commits January 9, 2026 09:35

text-to-speech mvp

a31a20c

add ios support

98dfda9

implement fallback phonemization (US)

b3a38b4

fix 'ed' phonemization bug

3f706d7

add british support

dc52f73

add cropping audio vector

1a98997

small refactor

0415505

demo app finished

d78c936

update model input variants

d65e87a

implement input partitioning

a38a1c1

native side streaming

6cfad82

audio streaming fixed

bb27b40

enable additional model options

8299362

fix reload bug

e4025b4

update phonemis binaries

18c5da3

reduce phonemis android binaries size & other small fixes

050cbc7

implement a demo quiz app

f9ff668

Merge branch '@is/text-to-speech' of https://github.com/software-mans…

fbe20dc

…ion/react-native-executorch into @is/text-to-speech

IgorSwat requested review from benITo47, chmjkb, mkopcins and msluszniak January 9, 2026 08:46

IgorSwat added 2 commits January 9, 2026 09:55

Remove Log.h include from RnExecutorchInstaller.h

60ac2e0

Removed unnecessary Log.h include from RnExecutorchInstaller.h

Remove unused iostream include

9f53a12

Removed unnecessary iostream include from BaseModel.cpp

msluszniak reviewed Jan 9, 2026

View reviewed changes

IgorSwat added 2 commits January 9, 2026 14:54

small code refactor

8a1db2a

Merge branch '@is/text-to-speech' of https://github.com/software-mans…

9eb0629

…ion/react-native-executorch into @is/text-to-speech

IgorSwat force-pushed the @is/text-to-speech branch from 571fada to 9eb0629 Compare January 9, 2026 13:58

chmjkb requested changes Jan 9, 2026

View reviewed changes

feat: text to speech (#546) #710

Are you sure you want to change the base?

feat: text to speech (#546) #710

Uh oh!

Conversation

IgorSwat commented Jan 9, 2026

Description

Introduces a breaking change?

Type of change

Tested on

Testing instructions

Screenshots

Related issues

Checklist

Additional notes

Uh oh!

msluszniak left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

msluszniak commented Jan 9, 2026

Uh oh!

chmjkb left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

chmjkb Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

chmjkb Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

chmjkb Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

chmjkb Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

chmjkb Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

chmjkb Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

chmjkb Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

msluszniak left a comment •

edited

Loading

chmjkb left a comment •

edited

Loading