openvinotoolkit · michalkulakowski · Jul 24, 2026 · Jul 24, 2026 · Jul 24, 2026 · Jul 24, 2026
diff --git a/demos/audio/README.md b/demos/audio/README.md
@@ -30,15 +30,15 @@ See the [T2s calculator documentation](../../docs/speech_generation/reference.md
 **Deploying with Docker**
 
 ```bash
-mkdir models
-docker run -d -u $(id -u):$(id -g) --rm -p 8000:8000 -v $(pwd)/models:/models:rw openvino/model_server:latest --rest_port 8000 --source_model luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_repository_path /models --model_name Kokoro-82M-OpenVINO-FP16-OVMS --target_device CPU --task text2speech
+mkdir -p ${HOME}/models
+docker run -d -u $(id -u):$(id -g) --rm -p 8000:8000 -v ${HOME}/models:/models:rw openvino/model_server:latest --rest_port 8000 --source_model luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_repository_path /models --model_name Kokoro-82M-OpenVINO-FP16-OVMS --target_device CPU
 ```
 
 **Deploying on Bare Metal**
 
 ```bat
 mkdir c:\models
-ovms --rest_port 8000 --source_model luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_repository_path c:\models --model_name Kokoro-82M-OpenVINO-FP16-OVMS --target_device CPU --task text2speech
+ovms --rest_port 8000 --source_model luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_repository_path c:\models --model_name Kokoro-82M-OpenVINO-FP16-OVMS --target_device CPU
 ```
 
 ### Request Generation 
@@ -107,33 +107,21 @@ Many variances of Whisper models can be deployed in a single command by using pr
 In this demo we will use OpenVINO/whisper-large-v3-turbo-fp16-ov, which is a fine-tuned version of Whisper large-v3.
 
 :::{dropdown} **Deploying with Docker**
-
-Select deployment option depending on how you prepared models in the previous step.
-
-**CPU**
-
-Running this command starts the container with CPU only target device:
-```bash
-mkdir -p models
-docker run -d -u $(id -u):$(id -g) --rm -p 8000:8000 -v $(pwd)/models:/models:rw openvino/model_server:latest --rest_port 8000 --task speech2text --source_model OpenVINO/whisper-large-v3-turbo-fp16-ov --model_name whisper-large-v3-turbo-fp16-ov --model_repository_path /models --target_device CPU
-```
-**GPU**
-
-In case you want to use GPU device to run the generation, add extra docker parameters `--device /dev/dri --group-add=$(stat -c "%g" /dev/dri/render* | head -n 1)`
-to `docker run` command, use the image with GPU support.
-It can be applied using the commands below:
 ```bash
-mkdir -p models
-docker run -d -u $(id -u):$(id -g) --rm -p 8000:8000 --device /dev/dri --group-add=$(stat -c "%g" /dev/dri/render* | head -n 1) -v $(pwd)/models:/models:rw openvino/model_server:latest-gpu --rest_port 8000 --task speech2text --source_model OpenVINO/whisper-large-v3-turbo-fp16-ov --model_name whisper-large-v3-turbo-fp16-ov --model_repository_path /models --target_device GPU
+mkdir -p ${HOME}/models
+# in case GPU is available
+export GPU_ARGS=$(if ls /dev/dri/render* >/dev/null 2>&1; then echo "--device /dev/dri --group-add $(stat -c '%g' /dev/dri/render* | head -n1)"; fi)
+docker run -d ${GPU_ARGS} -u $(id -u):$(id -g) --rm -p 8000:8000 -v ${HOME}/models:/models:rw openvino/model_server:weekly --rest_port 8000 --source_model OpenVINO/whisper-large-v3-turbo-fp16-ov --model_name whisper-large-v3-turbo-fp16-ov --model_repository_path /models
 ```
 :::
 
 :::{dropdown} **Deploying on Bare Metal**
 
-If you run on GPU make sure to have appropriate drivers installed, so the device is accessible for the model server.
+The same command can be used for CPU and GPU deployments. OVMS will auto-detect device settings when `--target_device` is not provided.
 
 ```bat
-ovms --rest_port 8000 --task speech2text --source_model OpenVINO/whisper-large-v3-turbo-fp16-ov --model_name whisper-large-v3-turbo-fp16-ov --model_repository_path models --target_device GPU
+mkdir c:\models
+ovms --rest_port 8000 --source_model OpenVINO/whisper-large-v3-turbo-fp16-ov --model_name whisper-large-v3-turbo-fp16-ov --model_repository_path c:\models
 ```
 :::
 
@@ -254,48 +242,29 @@ Prepare export script and dependencies:
 ```console
 curl https://raw.githubusercontent.com/openvinotoolkit/model_server/refs/heads/main/demos/common/export_models/export_model.py -o export_model.py
 pip install -r https://raw.githubusercontent.com/openvinotoolkit/model_server/refs/heads/main/demos/common/export_models/requirements.txt
-mkdir -p models
+mkdir -p ${HOME}/models
 ```
 
 Export Speech-to-Text model with word timestamps enabled.
-
-**CPU**
-
 ```console
-python export_model.py speech2text --source_model openai/whisper-large-v3-turbo --weight-format fp16 --model_name whisper-large-v3-turbo-word-ts --config_file_path models/config.json --model_repository_path models --overwrite_models --enable_word_timestamps --target_device CPU
-```
-
-**GPU**
-
-```console
-python export_model.py speech2text --source_model openai/whisper-large-v3-turbo --weight-format fp16 --model_name whisper-large-v3-turbo-word-ts --config_file_path models/config.json --model_repository_path models --overwrite_models --enable_word_timestamps --target_device GPU
+python export_model.py speech2text --source_model openai/whisper-large-v3-turbo --weight-format fp16 --model_name whisper-large-v3-turbo-word-ts --config_file_path ${HOME}/models/config.json --model_repository_path ${HOME}/models --overwrite_models --enable_word_timestamps
 ```
 
 :::{dropdown} **Deploying with Docker**
-
-Select deployment option depending on how you prepared models in the previous step.
-
-**CPU**
-
-```bash
-docker run -d -u $(id -u):$(id -g) --rm -p 8000:8000 -v $(pwd)/models:/models:rw openvino/model_server:latest --rest_port 8000 --config_path /models/config.json
-```
-
-**GPU**
-
-In case you want to use GPU device to run the generation, add extra docker parameters `--device /dev/dri --group-add=$(stat -c "%g" /dev/dri/render* | head -n 1)`
-to `docker run` command, use the image with GPU support.
-It can be applied using the commands below:
-
 ```bash
-docker run -d -u $(id -u):$(id -g) --rm -p 8000:8000 --device /dev/dri --group-add=$(stat -c "%g" /dev/dri/render* | head -n 1) -v $(pwd)/models:/models:rw openvino/model_server:latest-gpu --rest_port 8000 --config_path /models/config.json
+# in case GPU is available
+export GPU_ARGS=$(if ls /dev/dri/render* >/dev/null 2>&1; then echo "--device /dev/dri --group-add $(stat -c '%g' /dev/dri/render* | head -n1)"; fi)
+docker run -d ${GPU_ARGS} -u $(id -u):$(id -g) --rm -p 8000:8000 -v ${HOME}/models:/models:rw openvino/model_server:weekly --rest_port 8000 --config_path /models/config.json
 ```
 :::
 
 :::{dropdown} **Deploying on Bare Metal**
 
+The same command can be used for CPU and GPU deployments.
+
 ```bat
-ovms --rest_port 8000 --config_path models/config.json
+mkdir c:\models
+ovms --rest_port 8000 --config_path c:\models\config.json
 ```
 :::
 
@@ -402,34 +371,20 @@ To prepare an audio file with speech in a language other than English, e.g. Span
 
 **Deploying with Docker**
 
-**CPU**
-
 ```bash
-mkdir -p models
-docker run -d -u $(id -u):$(id -g) --rm -p 8000:8000 -v $(pwd)/models:/models:rw openvino/model_server:latest --rest_port 8000 --source_model luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_repository_path /models --model_name Kokoro-82M-OpenVINO-FP16-OVMS --target_device CPU --task text2speech
-```
-
-**GPU**
-
-```bash
-mkdir -p models
-docker run -d -u $(id -u):$(id -g) --rm -p 8000:8000 --device /dev/dri --group-add=$(stat -c "%g" /dev/dri/render* | head -n 1) -v $(pwd)/models:/models:rw openvino/model_server:latest-gpu --rest_port 8000 --source_model luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_repository_path /models --model_name Kokoro-82M-OpenVINO-FP16-OVMS --target_device GPU --task text2speech
+mkdir -p ${HOME}/models
+# in case GPU is available
+export GPU_ARGS=$(if ls /dev/dri/render* >/dev/null 2>&1; then echo "--device /dev/dri --group-add $(stat -c '%g' /dev/dri/render* | head -n1)"; fi)
+docker run -d ${GPU_ARGS} -u $(id -u):$(id -g) --rm -p 8000:8000 -v ${HOME}/models:/models:rw openvino/model_server:weekly --rest_port 8000 --source_model luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_repository_path /models --model_name Kokoro-82M-OpenVINO-FP16-OVMS
 ```
 
 **Deploying on Bare Metal**
 
-**CPU**
-
-```bat
-mkdir models
-ovms --rest_port 8000 --source_model luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_repository_path models --model_name Kokoro-82M-OpenVINO-FP16-OVMS --target_device CPU --task text2speech
-```
-
-**GPU**
+The same command can be used for CPU and GPU deployments. OVMS will auto-detect device settings when `--target_device` is not provided.
 
 ```bat
-mkdir models
-ovms --rest_port 8000 --source_model luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_repository_path models --model_name Kokoro-82M-OpenVINO-FP16-OVMS --target_device GPU --task text2speech
+mkdir c:\models
+ovms --rest_port 8000 --source_model luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_repository_path c:\models --model_name Kokoro-82M-OpenVINO-FP16-OVMS
 ```
 
 For non-English Kokoro input, set the `language` field explicitly.
@@ -443,38 +398,21 @@ Whisper models can be deployed in a single command by using pre-configured model
 Here is an example of OpenVINO/whisper-large-v3-fp16-ov deployment:
 
 :::{dropdown} **Deploying with Docker**
-
-Select deployment option depending on how you prepared models in the previous step.
-
-**CPU**
-
-Running this command starts the container with CPU only target device:
-```bash
-mkdir -p models
-docker run -d -u $(id -u):$(id -g) --rm -p 8000:8000 -v $(pwd)/models:/models:rw openvino/model_server:latest --rest_port 8000 --source_model OpenVINO/whisper-large-v3-fp16-ov --model_repository_path /models --model_name whisper-large-v3-fp16-ov --task speech2text --target_device CPU
-```
-**GPU**
-
-In case you want to use GPU device to run the generation, add extra docker parameters `--device /dev/dri --group-add=$(stat -c "%g" /dev/dri/render* | head -n 1)`
-to `docker run` command, use the image with GPU support.
-It can be applied using the commands below:
 ```bash
-mkdir -p models
-docker run -d -u $(id -u):$(id -g) --rm -p 8000:8000 --device /dev/dri --group-add=$(stat -c "%g" /dev/dri/render* | head -n 1) -v $(pwd)/models:/models:rw openvino/model_server:latest-gpu --rest_port 8000 --source_model OpenVINO/whisper-large-v3-fp16-ov --model_repository_path /models --model_name whisper-large-v3-fp16-ov --task speech2text --target_device GPU
+mkdir -p ${HOME}/models
+# in case GPU is available
+export GPU_ARGS=$(if ls /dev/dri/render* >/dev/null 2>&1; then echo "--device /dev/dri --group-add $(stat -c '%g' /dev/dri/render* | head -n1)"; fi)
+docker run -d ${GPU_ARGS} -u $(id -u):$(id -g) --rm -p 8000:8000 -v ${HOME}/models:/models:rw openvino/model_server:weekly --rest_port 8000 --source_model OpenVINO/whisper-large-v3-fp16-ov --model_repository_path /models --model_name whisper-large-v3-fp16-ov
 ```
 :::
 
 :::{dropdown} **Deploying on Bare Metal**
 
-If you run on GPU make sure to have appropriate drivers installed, so the device is accessible for the model server.
+The same command can be used for CPU and GPU deployments. OVMS will auto-detect device settings when `--target_device` is not provided.
 
 ```bat
-mkdir models
-ovms --rest_port 8000 --source_model OpenVINO/whisper-large-v3-fp16-ov --model_repository_path models --model_name whisper-large-v3-fp16-ov --task speech2text --target_device CPU
-```
-or
-```bat
-ovms --rest_port 8000 --source_model OpenVINO/whisper-large-v3-fp16-ov --model_repository_path models --model_name whisper-large-v3-fp16-ov --task speech2text --target_device GPU
+mkdir c:\models
+ovms --rest_port 8000 --source_model OpenVINO/whisper-large-v3-fp16-ov --model_repository_path c:\models --model_name whisper-large-v3-fp16-ov
 ```
 :::
 

diff --git a/demos/integration_with_OpenWebUI/README.md b/demos/integration_with_OpenWebUI/README.md
@@ -461,11 +461,10 @@ Then it's ready to use. In new chat it's possible to toggle **Code Interpreter**
 
 ### Step 1: Models Preparation
 
-Start by downloading `export_model.py` script and run it to download and quantize the model for speech generation:
+Start by pulling the pre-exported Kokoro model with OVMS and adding it to the server config:
 ```console 
-curl https://raw.githubusercontent.com/openvinotoolkit/model_server/refs/heads/main/demos/common/export_models/export_model.py -o export_model.py
-pip3 install -r https://raw.githubusercontent.com/openvinotoolkit/model_server/refs/heads/main/demos/common/export_models/requirements.txt
-python export_model.py text2speech --source_model microsoft/speecht5_tts --weight-format fp32 --model_name microsoft/speecht5_tts --config_file_path models/config.json --model_repository_path models --vocoder microsoft/speecht5_hifigan
+docker run --rm -u $(id -u):$(id -g) -v $PWD/models:/models --device /dev/dri --group-add=$(stat -c "%g" /dev/dri/render* | head -n 1) openvino/model_server:weekly --pull --source_model luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_repository_path /models --target_device GPU
+docker run --rm -u $(id -u):$(id -g) -v $PWD/models:/models openvino/model_server:weekly --add_to_config --config_path /models/config.json --model_path luis-castillo/Kokoro-82M-OpenVINO-FP16-OVMS --model_name Kokoro-82M-OpenVINO-FP16-OVMS
 ```
 
 Next, download and add to config model for transcription:
@@ -496,7 +495,7 @@ docker run --rm -u $(id -u):$(id -g) -v $PWD/models:/models openvino/model_serve
    * URL: `http://localhost:8000/v3`
    * Set Engine type to `OpenAI` 
    * STT Model: `OpenVINO/whisper-base-fp16-ov`
-   * TTS Model: `microsoft/speecht5_tts`
+   * TTS Model: `Kokoro-82M-OpenVINO-FP16-OVMS`
    * Put anything in API key
 3. Click **Save**
 

diff --git a/docs/clients_genai.md b/docs/clients_genai.md
@@ -463,8 +463,8 @@ speech_file_path = Path(__file__).parent / "speech.wav"
 client = OpenAI(base_url=url, api_key="not_used")
 
 with client.audio.speech.with_streaming_response.create(
-  model="microsoft/speecht5_tts",
-  voice="unused",
+  model="Kokoro-82M-OpenVINO-FP16-OVMS",
+  voice="af_alloy",
   input=prompt
 ) as response:
   response.stream_to_file(speech_file_path)
@@ -475,7 +475,7 @@ with client.audio.speech.with_streaming_response.create(
 ```text
 curl http://localhost:8000/v3/audio/speech \
   -H "Content-Type: application/json" \
-  -d "{\"model\": \"microsoft/speecht5_tts\", \"input\": \"The quick brown fox jumped over the lazy dog\"}" \
+  -d "{\"model\": \"Kokoro-82M-OpenVINO-FP16-OVMS\", \"voice\": \"af_alloy\", \"input\": \"The quick brown fox jumped over the lazy dog\"}" \
   -o speech.wav
 ```
 :::

diff --git a/docs/genai.md b/docs/genai.md
@@ -57,7 +57,7 @@ Check also the demo of [transcription](../demos/audio/README.md#transcription) a
 ## Speech generation
 
 The implementation is compatible with the OpenAI API for [audio/speech](./model_server_rest_api_text_to_speech.md).
-It supports microsoft/speecht5_tts model.
+It supports Kokoro speech generation models, for example `Kokoro-82M-OpenVINO-FP16-OVMS`.
 
 Check also the demo of [speech generation](../demos/audio/README.md#speech-generation).
 

diff --git a/docs/model_server_rest_api_text_to_speech.md b/docs/model_server_rest_api_text_to_speech.md
@@ -16,8 +16,9 @@ Request body must be in JSON format, and the request must have `Content-Type: ap
 curl http://localhost:8000/v3/audio/speech \
   -H "Content-Type: application/json" \
   -d '{
-    "model": "microsoft/speecht5_tts",
-    "input": "The quick brown fox jumped over the lazy dog.",
+    "model": "Kokoro-82M-OpenVINO-FP16-OVMS",
+    "voice": "af_alloy",
+    "input": "The quick brown fox jumped over the lazy dog."
   }' \
   -o speech.wav
 ```

diff --git a/docs/parameters.md b/docs/parameters.md
@@ -118,7 +118,6 @@ When pulling models outside of OpenVINO organization the optimum-cli api is used
 |------------------------------|--------------|---------------------------------------------------------------------------------------------------------------|
 | `--extra_quantization_params`| `string`     | Add advanced quantization parameters. Check [optimum-intel](https://github.com/huggingface/optimum-intel) documentation. Example: `--sym --group-size -1 --ratio 1.0 --awq --scale-estimation --dataset wikitext2`  |
 | `--weight-format`            | `string`     | Model precision used in optimum-cli export with conversion. Default `int8`.                |
-| `--vocoder`                  | `string`     | The vocoder model to use for text2speech. For example `microsoft/speecht5_hifigan`.        |
 
 There are also additional environment variables that may change the behavior of pulling:
 
@@ -190,8 +189,7 @@ Task specific parameters for different tasks (text generation/image generation/e
 | option                    | Value format | Description                                                                    |
 |---------------------------|--------------|--------------------------------------------------------------------------------|
 | `--num_streams`           | `integer`    | The number of parallel execution streams to use for the model. Use at least 2 on 2 socket CPU systems. Default: 1. |
-| `--model_type`            | `string`     | Type of the source TTS model: `speecht5` (default) or `kokoro`.                |
-| `--vocoder`               | `string`     | The vocoder model to use for text2speech. For example `microsoft/speecht5_hifigan`. |
+| `--model_type`            | `string`     | Type of the source TTS model. Supported values: `kokoro` (default, recommended) and `speecht5` (legacy). |
 
 ### Speech to text
 | option                    | Value format | Description                                                                    |