MCP

README.md

@@ -4,7 +4,7 @@ emoji: 🔥
 colorFrom: yellow
 colorTo: red
 sdk: gradio
-sdk_version: 5.16.0
+sdk_version: 5.31.0
 app_file: app.py
 pinned: false
 ---

app.py

@@ -1,5 +1,12 @@
 #!/usr/bin/env python
 
+import os
+import shlex
+import subprocess
+
+if os.getenv("SPACE_ID"):
+    subprocess.run(shlex.split("python -m unidic download"), timeout=100, check=True)  # noqa: S603
+
 import gradio as gr
 import numpy as np
 import spaces
@@ -17,7 +24,24 @@ VOICES = [
 
 
 @spaces.GPU(duration=20)
-def run(text: str, voice: str, speed: float) -> tuple[tuple[int, np.ndarray], str]:
+def run(text: str, voice: str, speed: float = 1.0) -> tuple[tuple[int, np.ndarray], str]:
+    """Generate audio from text using Kokoro.
+
+    Available voices are:
+        - jf_alpha
+        - jf_gongitsune
+        - jf_nezumi
+        - jf_tebukuro
+        - jm_kumo
+
+    Args:
+        text (str): Text to generate audio from.
+        voice (str): Voice to use.
+        speed (float): Speed of the audio. Defaults to 1.0.
+
+    Returns:
+        tuple[tuple[int, np.ndarray], str]: Tuple of (sample rate, audio data) and the text.
+    """
     generator = pipeline(
         text,
         voice=voice,
@@ -58,4 +82,4 @@ with gr.Blocks(css_paths="style.css") as demo:
     )
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch(mcp_server=True)

pyproject.toml

@@ -5,12 +5,12 @@ description = ""
 readme = "README.md"
 requires-python = ">=3.10"
 dependencies = [
-    "gradio>=5.16.0",
+    "gradio[mcp]>=5.31.0",
     "hf-transfer>=0.1.9",
-    "kokoro>=0.7.16",
-    "misaki[ja]>=0.7.17",
-    "spaces>=0.32.0",
-    "torch==2.4.0",
+    "kokoro>=0.9.4",
+    "misaki[ja]>=0.9.4",
+    "spaces>=0.36.0",
+    "torch==2.5.1",
 ]
 
 [tool.ruff]

requirements.txt

@@ -1,5 +1,7 @@
 # This file was autogenerated by uv via the following command:
 #    uv pip compile pyproject.toml -o requirements.txt
+addict==2.4.0
+    # via misaki
 aiofiles==23.2.1
     # via gradio
 annotated-types==0.7.0
@@ -8,6 +10,8 @@ anyio==4.8.0
     # via
     #   gradio
     #   httpx
+    #   mcp
+    #   sse-starlette
     #   starlette
 attrs==25.1.0
     # via
@@ -84,11 +88,13 @@ fsspec==2025.2.0
     #   torch
 fugashi==1.4.0
     # via misaki
-gradio==5.16.0
+gradio==5.31.0
     # via
     #   kokoro-ja (pyproject.toml)
     #   spaces
-gradio-client==1.7.0
+gradio-client==1.10.1
+    # via gradio
+groovy==0.1.2
     # via gradio
 h11==0.14.0
     # via
@@ -102,8 +108,11 @@ httpx==0.28.1
     # via
     #   gradio
     #   gradio-client
+    #   mcp
     #   safehttpx
     #   spaces
+httpx-sse==0.4.0
+    # via mcp
 huggingface-hub==0.28.1
     # via
     #   gradio
@@ -133,7 +142,7 @@ jsonschema==4.23.0
     # via csvw
 jsonschema-specifications==2024.10.1
     # via jsonschema
-kokoro==0.7.16
+kokoro==0.9.4
     # via kokoro-ja (pyproject.toml)
 langcodes==3.5.0
     # via spacy
@@ -156,9 +165,11 @@ markupsafe==2.1.5
     #   clldutils
     #   gradio
     #   jinja2
+mcp==1.9.0
+    # via gradio
 mdurl==0.1.2
     # via markdown-it-py
-misaki==0.7.17
+misaki==0.9.4
     # via
     #   kokoro-ja (pyproject.toml)
     #   kokoro
@@ -181,40 +192,41 @@ numpy==1.26.4
     #   gradio
     #   kokoro
     #   pandas
-    #   scipy
+    #   pyopenjtalk
     #   spacy
     #   thinc
     #   transformers
-nvidia-cublas-cu12==12.1.3.1
+nvidia-cublas-cu12==12.4.5.8
     # via
     #   nvidia-cudnn-cu12
     #   nvidia-cusolver-cu12
     #   torch
-nvidia-cuda-cupti-cu12==12.1.105
+nvidia-cuda-cupti-cu12==12.4.127
     # via torch
-nvidia-cuda-nvrtc-cu12==12.1.105
+nvidia-cuda-nvrtc-cu12==12.4.127
     # via torch
-nvidia-cuda-runtime-cu12==12.1.105
+nvidia-cuda-runtime-cu12==12.4.127
     # via torch
 nvidia-cudnn-cu12==9.1.0.70
     # via torch
-nvidia-cufft-cu12==11.0.2.54
+nvidia-cufft-cu12==11.2.1.3
     # via torch
-nvidia-curand-cu12==10.3.2.106
+nvidia-curand-cu12==10.3.5.147
     # via torch
-nvidia-cusolver-cu12==11.4.5.107
+nvidia-cusolver-cu12==11.6.1.9
     # via torch
-nvidia-cusparse-cu12==12.1.0.106
+nvidia-cusparse-cu12==12.3.1.170
     # via
     #   nvidia-cusolver-cu12
     #   torch
-nvidia-nccl-cu12==2.20.5
+nvidia-nccl-cu12==2.21.5
     # via torch
-nvidia-nvjitlink-cu12==12.8.61
+nvidia-nvjitlink-cu12==12.4.127
     # via
     #   nvidia-cusolver-cu12
     #   nvidia-cusparse-cu12
-nvidia-nvtx-cu12==12.1.105
+    #   torch
+nvidia-nvtx-cu12==12.4.127
     # via torch
 orjson==3.10.15
     # via gradio
@@ -234,29 +246,37 @@ phonemizer-fork==3.3.2
     # via misaki
 pillow==11.1.0
     # via gradio
+plac==1.4.5
+    # via unidic
 preshed==3.0.9
     # via
     #   spacy
     #   thinc
 psutil==5.9.8
     # via spaces
-pydantic==2.10.6
+pydantic==2.11.5
     # via
     #   confection
     #   fastapi
     #   gradio
+    #   mcp
+    #   pydantic-settings
     #   spaces
     #   spacy
     #   thinc
     #   weasel
-pydantic-core==2.27.2
+pydantic-core==2.33.2
     # via pydantic
+pydantic-settings==2.9.1
+    # via mcp
 pydub==0.25.1
     # via gradio
 pygments==2.19.1
     # via rich
 pylatexenc==2.10
     # via clldutils
+pyopenjtalk==0.4.1
+    # via misaki
 pyparsing==3.2.1
     # via rdflib
 python-dateutil==2.9.0.post0
@@ -264,8 +284,12 @@ python-dateutil==2.9.0.post0
     #   clldutils
     #   csvw
     #   pandas
+python-dotenv==1.1.0
+    # via pydantic-settings
 python-multipart==0.0.20
-    # via gradio
+    # via
+    #   gradio
+    #   mcp
 pytz==2025.1
     # via pandas
 pyyaml==6.0.2
@@ -292,6 +316,7 @@ requests==2.32.3
     #   spaces
     #   spacy
     #   transformers
+    #   unidic
     #   weasel
 rfc3986==1.5.0
     # via csvw
@@ -307,8 +332,6 @@ safehttpx==0.1.6
     # via gradio
 safetensors==0.5.2
     # via transformers
-scipy==1.15.2
-    # via kokoro
 segments==2.2.1
     # via phonemizer-fork
 semantic-version==2.10.0
@@ -326,7 +349,7 @@ smart-open==7.1.0
     # via weasel
 sniffio==1.3.1
     # via anyio
-spaces==0.32.0
+spaces==0.36.0
     # via kokoro-ja (pyproject.toml)
 spacy==3.8.4
     # via misaki
@@ -342,11 +365,15 @@ srsly==2.5.1
     #   spacy
     #   thinc
     #   weasel
+sse-starlette==2.3.5
+    # via mcp
 starlette==0.45.3
     # via
     #   fastapi
     #   gradio
-sympy==1.13.3
+    #   mcp
+    #   sse-starlette
+sympy==1.13.1
     # via torch
 tabulate==0.9.0
     # via clldutils
@@ -356,7 +383,7 @@ tokenizers==0.21.0
     # via transformers
 tomlkit==0.13.2
     # via gradio
-torch==2.4.0
+torch==2.5.1
     # via
     #   kokoro-ja (pyproject.toml)
     #   curated-transformers
@@ -365,11 +392,13 @@ torch==2.4.0
 tqdm==4.67.1
     # via
     #   huggingface-hub
+    #   pyopenjtalk
     #   spacy
     #   transformers
+    #   unidic
 transformers==4.48.3
     # via kokoro
-triton==3.0.0
+triton==3.1.0
     # via torch
 typer==0.15.1
     # via
@@ -392,21 +421,29 @@ typing-extensions==4.12.2
     #   spaces
     #   torch
     #   typer
+    #   typing-inspection
     #   uvicorn
+typing-inspection==0.4.1
+    # via
+    #   pydantic
+    #   pydantic-settings
 tzdata==2025.1
     # via pandas
-unidic-lite==1.0.8
+unidic==1.1.0
     # via misaki
 uritemplate==4.1.1
     # via csvw
 urllib3==2.3.0
     # via requests
 uvicorn==0.34.0
-    # via gradio
-wasabi==1.1.3
+    # via
+    #   gradio
+    #   mcp
+wasabi==0.10.1
     # via
     #   spacy
     #   thinc
+    #   unidic
     #   weasel
 weasel==0.4.1
     # via spacy