mcp-video2text
3.2
If you are the rightful owner of mcp-video2text and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
mcp-video2text is a tool designed to convert online videos into text, allowing users to quickly read the content of videos.
mcp-video2text
Convert online video to text, read it quickly
Usage
## require
$ curl -LsSf https://astral.sh/uv/install.sh | sh # install uv package manager
$ brew install ffmpeg # or `sudo apt install ffmpeg` for linux
## install
$ uv pip install mcp-video2text[mlx] # recommend with mlx for silicon machine
$ uv pip install mcp-video2text[cpp]
for mcp client, pass uv run mpc-video2text to command field.
{
"mcpServers": {
"mcp-video2text": {
"command": "uv",
"args": ["run", "mcp-video2text"],
}
}
}
The first run will download the Whisper model and may take a few minutes.
It's recommended to update the timeout settings in your client.
or your can custom your own client, view example test_mcp_client.py
import asyncio
import os
from mcp import ClientSession, StdioServerParameters
from mcp.client.stdio import stdio_client
async def client():
server_params = StdioServerParameters(
command="uv", args=["run", "mcp-video2text"], env=os.environ
)
async with stdio_client(server_params) as (read, write):
async with ClientSession(read, write) as session:
await session.initialize()
print(await session.list_tools())
result = await session.call_tool(
"video2text", {"url": "https://www.bilibili.com/video/BV1gdERzuEYB/"}
)
print(result.content[0].text)
if __name__ == "__main__":
asyncio.run(client())
