Add C API for Kokoro TTS models (#1717)

.github/workflows/c-api.yaml

@@ -79,6 +79,32 @@ jobs:
             otool -L ./install/lib/libsherpa-onnx-c-api.dylib
           fi
 
+      - name: Test Kokoro TTS (en)
+        shell: bash
+        run: |
+          gcc -o kokoro-tts-en-c-api ./c-api-examples/kokoro-tts-en-c-api.c \
+            -I ./build/install/include \
+            -L ./build/install/lib/ \
+            -l sherpa-onnx-c-api \
+            -l onnxruntime
+
+          curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/kokoro-en-v0_19.tar.bz2
+          tar xf kokoro-en-v0_19.tar.bz2
+          rm kokoro-en-v0_19.tar.bz2
+
+          export LD_LIBRARY_PATH=$PWD/build/install/lib:$LD_LIBRARY_PATH
+          export DYLD_LIBRARY_PATH=$PWD/build/install/lib:$DYLD_LIBRARY_PATH
+
+          ./kokoro-tts-en-c-api
+
+          rm ./kokoro-tts-en-c-api
+          rm -rf kokoro-en-*
+
+      - uses: actions/upload-artifact@v4
+        with:
+          name: kokoro-tts-${{ matrix.os }}
+          path: ./generated-kokoro-*.wav
+
       - name: Test Matcha TTS (zh)
         shell: bash
         run: |

.github/workflows/cxx-api.yaml

@@ -81,6 +81,33 @@ jobs:
             otool -L ./install/lib/libsherpa-onnx-cxx-api.dylib
           fi
 
+      - name: Test Kokoro TTS (en)
+        shell: bash
+        run: |
+          g++ -std=c++17 -o kokoro-tts-en-cxx-api ./cxx-api-examples/kokoro-tts-en-cxx-api.cc \
+            -I ./build/install/include \
+            -L ./build/install/lib/ \
+            -l sherpa-onnx-cxx-api \
+            -l sherpa-onnx-c-api \
+            -l onnxruntime
+
+          curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/kokoro-en-v0_19.tar.bz2
+          tar xf kokoro-en-v0_19.tar.bz2
+          rm kokoro-en-v0_19.tar.bz2
+
+          export LD_LIBRARY_PATH=$PWD/build/install/lib:$LD_LIBRARY_PATH
+          export DYLD_LIBRARY_PATH=$PWD/build/install/lib:$DYLD_LIBRARY_PATH
+
+          ./kokoro-tts-en-cxx-api
+
+          rm kokoro-tts-en-cxx-api
+          rm -rf kokoro-en-*
+
+      - uses: actions/upload-artifact@v4
+        with:
+          name: kokoro-tts-${{ matrix.os }}
+          path: ./generated-kokoro-*.wav
+
       - name: Test Matcha TTS (zh)
         shell: bash
         run: |

.gitignore

@@ -127,3 +127,4 @@ harmony-os/SherpaOnnxHar/sherpa_onnx/LICENSE
 harmony-os/SherpaOnnxHar/sherpa_onnx/CHANGELOG.md
 matcha-icefall-zh-baker
 matcha-icefall-en_US-ljspeech
+kokoro-en-v0_19

c-api-examples/CMakeLists.txt

@@ -13,6 +13,9 @@ if(SHERPA_ONNX_ENABLE_TTS)
 
   add_executable(matcha-tts-en-c-api matcha-tts-en-c-api.c)
   target_link_libraries(matcha-tts-en-c-api sherpa-onnx-c-api)
+
+  add_executable(kokoro-tts-en-c-api kokoro-tts-en-c-api.c)
+  target_link_libraries(kokoro-tts-en-c-api sherpa-onnx-c-api)
 endif()
 
 if(SHERPA_ONNX_ENABLE_SPEAKER_DIARIZATION)

c-api-examples/kokoro-tts-en-c-api.c

@@ -0,0 +1,84 @@
+// c-api-examples/kokoro-tts-en-c-api.c
+//
+// Copyright (c)  2025  Xiaomi Corporation
+
+// This file shows how to use sherpa-onnx C API
+// for English TTS with Kokoro.
+//
+// clang-format off
+/*
+Usage
+
+
+wget https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/kokoro-en-v0_19.tar.bz2
+tar xf kokoro-en-v0_19.tar.bz2
+rm kokoro-en-v0_19.tar.bz2
+
+./kokoro-tts-en-c-api
+
+ */
+// clang-format on
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "sherpa-onnx/c-api/c-api.h"
+
+static int32_t ProgressCallback(const float *samples, int32_t num_samples,
+                                float progress) {
+  fprintf(stderr, "Progress: %.3f%%\n", progress * 100);
+  // return 1 to continue generating
+  // return 0 to stop generating
+  return 1;
+}
+
+int32_t main(int32_t argc, char *argv[]) {
+  SherpaOnnxOfflineTtsConfig config;
+  memset(&config, 0, sizeof(config));
+  config.model.kokoro.model = "./kokoro-en-v0_19/model.onnx";
+  config.model.kokoro.voices = "./kokoro-en-v0_19/voices.bin";
+  config.model.kokoro.tokens = "./kokoro-en-v0_19/tokens.txt";
+  config.model.kokoro.data_dir = "./kokoro-en-v0_19/espeak-ng-data";
+
+  config.model.num_threads = 2;
+
+  // If you don't want to see debug messages, please set it to 0
+  config.model.debug = 1;
+
+  const char *filename = "./generated-kokoro-en.wav";
+  const char *text =
+      "Today as always, men fall into two groups: slaves and free men. Whoever "
+      "does not have two-thirds of his day for himself, is a slave, whatever "
+      "he may be: a statesman, a businessman, an official, or a scholar. "
+      "Friends fell out often because life was changing so fast. The easiest "
+      "thing in the world was to lose touch with someone.";
+
+  const SherpaOnnxOfflineTts *tts = SherpaOnnxCreateOfflineTts(&config);
+  // mapping of sid to voice name
+  // 0->af, 1->af_bella, 2->af_nicole, 3->af_sarah, 4->af_sky, 5->am_adam
+  // 6->am_michael, 7->bf_emma, 8->bf_isabella, 9->bm_george, 10->bm_lewis
+  int32_t sid = 0;
+  float speed = 1.0;  // larger -> faster in speech speed
+
+#if 0
+  // If you don't want to use a callback, then please enable this branch
+  const SherpaOnnxGeneratedAudio *audio =
+      SherpaOnnxOfflineTtsGenerate(tts, text, sid, speed);
+#else
+  const SherpaOnnxGeneratedAudio *audio =
+      SherpaOnnxOfflineTtsGenerateWithProgressCallback(tts, text, sid, speed,
+                                                       ProgressCallback);
+#endif
+
+  SherpaOnnxWriteWave(audio->samples, audio->n, audio->sample_rate, filename);
+
+  SherpaOnnxDestroyOfflineTtsGeneratedAudio(audio);
+  SherpaOnnxDestroyOfflineTts(tts);
+
+  fprintf(stderr, "Input text is: %s\n", text);
+  fprintf(stderr, "Speaker ID is is: %d\n", sid);
+  fprintf(stderr, "Saved to: %s\n", filename);
+
+  return 0;
+}

cxx-api-examples/CMakeLists.txt

@@ -21,4 +21,7 @@ if(SHERPA_ONNX_ENABLE_TTS)
 
   add_executable(matcha-tts-en-cxx-api ./matcha-tts-en-cxx-api.cc)
   target_link_libraries(matcha-tts-en-cxx-api sherpa-onnx-cxx-api)
+
+  add_executable(kokoro-tts-en-cxx-api ./kokoro-tts-en-cxx-api.cc)
+  target_link_libraries(kokoro-tts-en-cxx-api sherpa-onnx-cxx-api)
 endif()

cxx-api-examples/kokoro-tts-en-cxx-api.cc

@@ -0,0 +1,73 @@
+// cxx-api-examples/kokoro-tts-en-cxx-api.c
+//
+// Copyright (c)  2025  Xiaomi Corporation
+
+// This file shows how to use sherpa-onnx CXX API
+// for Chinese TTS with Kokoro.
+//
+// clang-format off
+/*
+Usage
+
+wget https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/kokoro-en-v0_19.tar.bz2
+tar xf kokoro-en-v0_19.tar.bz2
+rm kokoro-en-v0_19.tar.bz2
+
+./kokoro-tts-en-cxx-api
+
+ */
+// clang-format on
+
+#include <string>
+
+#include "sherpa-onnx/c-api/cxx-api.h"
+
+static int32_t ProgressCallback(const float *samples, int32_t num_samples,
+                                float progress, void *arg) {
+  fprintf(stderr, "Progress: %.3f%%\n", progress * 100);
+  // return 1 to continue generating
+  // return 0 to stop generating
+  return 1;
+}
+
+int32_t main(int32_t argc, char *argv[]) {
+  using namespace sherpa_onnx::cxx;  // NOLINT
+  OfflineTtsConfig config;
+
+  config.model.kokoro.model = "./kokoro-en-v0_19/model.onnx";
+  config.model.kokoro.voices = "./kokoro-en-v0_19/voices.bin";
+  config.model.kokoro.tokens = "./kokoro-en-v0_19/tokens.txt";
+  config.model.kokoro.data_dir = "./kokoro-en-v0_19/espeak-ng-data";
+
+  config.model.num_threads = 2;
+
+  // If you don't want to see debug messages, please set it to 0
+  config.model.debug = 1;
+
+  std::string filename = "./generated-kokoro-en-cxx.wav";
+  std::string text =
+      "Today as always, men fall into two groups: slaves and free men. Whoever "
+      "does not have two-thirds of his day for himself, is a slave, whatever "
+      "he may be: a statesman, a businessman, an official, or a scholar. "
+      "Friends fell out often because life was changing so fast. The easiest "
+      "thing in the world was to lose touch with someone.";
+
+  auto tts = OfflineTts::Create(config);
+  int32_t sid = 0;
+  float speed = 1.0;  // larger -> faster in speech speed
+
+#if 0
+  // If you don't want to use a callback, then please enable this branch
+  GeneratedAudio audio = tts.Generate(text, sid, speed);
+#else
+  GeneratedAudio audio = tts.Generate(text, sid, speed, ProgressCallback);
+#endif
+
+  WriteWave(filename, {audio.samples, audio.sample_rate});
+
+  fprintf(stderr, "Input text is: %s\n", text.c_str());
+  fprintf(stderr, "Speaker ID is is: %d\n", sid);
+  fprintf(stderr, "Saved to: %s\n", filename.c_str());
+
+  return 0;
+}

sherpa-onnx/c-api/c-api.cc

@@ -1092,6 +1092,18 @@ static sherpa_onnx::OfflineTtsConfig GetOfflineTtsConfig(
   tts_config.model.matcha.dict_dir =
       SHERPA_ONNX_OR(config->model.matcha.dict_dir, "");
 
+  // kokoro
+  tts_config.model.kokoro.model =
+      SHERPA_ONNX_OR(config->model.kokoro.model, "");
+  tts_config.model.kokoro.voices =
+      SHERPA_ONNX_OR(config->model.kokoro.voices, "");
+  tts_config.model.kokoro.tokens =
+      SHERPA_ONNX_OR(config->model.kokoro.tokens, "");
+  tts_config.model.kokoro.data_dir =
+      SHERPA_ONNX_OR(config->model.kokoro.data_dir, "");
+  tts_config.model.kokoro.length_scale =
+      SHERPA_ONNX_OR(config->model.kokoro.length_scale, 1.0);
+
   tts_config.model.num_threads = SHERPA_ONNX_OR(config->model.num_threads, 1);
   tts_config.model.debug = config->model.debug;
   tts_config.model.provider = SHERPA_ONNX_OR(config->model.provider, "cpu");

sherpa-onnx/c-api/c-api.h

@@ -910,12 +910,22 @@ SHERPA_ONNX_API typedef struct SherpaOnnxOfflineTtsMatchaModelConfig {
   const char *dict_dir;
 } SherpaOnnxOfflineTtsMatchaModelConfig;
 
+SHERPA_ONNX_API typedef struct SherpaOnnxOfflineTtsKokoroModelConfig {
+  const char *model;
+  const char *voices;
+  const char *tokens;
+  const char *data_dir;
+
+  float length_scale;  // < 1, faster in speech speed; > 1, slower in speed
+} SherpaOnnxOfflineTtsKokoroModelConfig;
+
 SHERPA_ONNX_API typedef struct SherpaOnnxOfflineTtsModelConfig {
   SherpaOnnxOfflineTtsVitsModelConfig vits;
   int32_t num_threads;
   int32_t debug;
   const char *provider;
   SherpaOnnxOfflineTtsMatchaModelConfig matcha;
+  SherpaOnnxOfflineTtsKokoroModelConfig kokoro;
 } SherpaOnnxOfflineTtsModelConfig;
 
 SHERPA_ONNX_API typedef struct SherpaOnnxOfflineTtsConfig {

sherpa-onnx/c-api/cxx-api.cc

@@ -338,6 +338,12 @@ OfflineTts OfflineTts::Create(const OfflineTtsConfig &config) {
   c.model.matcha.length_scale = config.model.matcha.length_scale;
   c.model.matcha.dict_dir = config.model.matcha.dict_dir.c_str();
 
+  c.model.kokoro.model = config.model.kokoro.model.c_str();
+  c.model.kokoro.voices = config.model.kokoro.voices.c_str();
+  c.model.kokoro.tokens = config.model.kokoro.tokens.c_str();
+  c.model.kokoro.data_dir = config.model.kokoro.data_dir.c_str();
+  c.model.kokoro.length_scale = config.model.kokoro.length_scale;
+
   c.model.num_threads = config.model.num_threads;
   c.model.debug = config.model.debug;
   c.model.provider = config.model.provider.c_str();

sherpa-onnx/c-api/cxx-api.h

@@ -338,9 +338,19 @@ struct OfflineTtsMatchaModelConfig {
   float length_scale = 1.0;  // < 1, faster in speed; > 1, slower in speed
 };
 
+struct OfflineTtsKokoroModelConfig {
+  std::string model;
+  std::string voices;
+  std::string tokens;
+  std::string data_dir;
+
+  float length_scale = 1.0;  // < 1, faster in speed; > 1, slower in speed
+};
+
 struct OfflineTtsModelConfig {
   OfflineTtsVitsModelConfig vits;
   OfflineTtsMatchaModelConfig matcha;
+  OfflineTtsKokoroModelConfig kokoro;
   int32_t num_threads = 1;
   bool debug = false;
   std::string provider = "cpu";