ai-translation-tool
Allows using NVIDIA's AI gateway (e.g., Nemotron models) as the backend for generating meeting minutes from transcripts.
Allows using local Ollama models to generate meeting minutes from transcripts, enabling offline or private processing.
Allows using OpenAI-compatible AI models to generate structured requirement research meeting minutes from transcribed audio.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@ai-translation-tool转写这段客户访谈录音并生成需求调研纪要"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
客户需求调研纪要工具
浏览器录音 / 上传音频 → Whisper 转写 → AI 生成结构化《需求调研纪要》(含业务流程图)→ 导出 md/html/pdf/docx。 支持 Web 页面、MCP 协议、命令行三种方式调用,可集成到 workbuddy 等工具。
一、安装依赖
要求 Python 3.10+(推荐 3.13)。国内网络建议先设置镜像:
pip install -r requirements.txt如需 GPU 转写更快,安装 CUDA 版 ctranslate2 并复制对应 cublas/cudnn DLL(见项目说明)。
Related MCP server: audio-transcription-mcp
二、配置
复制 .env.example 为 .env 并填写:
copy .env.example .env配置项 | 说明 |
| AI 后端: |
| OpenAI 兼容后端的 API Key;Ollama 无需填写 |
| OpenAI 兼容后端地址,默认 NVIDIA 网关 |
| 模型名,默认 NVIDIA Nemotron |
| 默认转写模型(内置 small 或自定义模型名) |
| 自定义 Whisper 本地模型(路径需存在) |
三、方式一:Web 页面
python app.py --port 5000浏览器打开 http://127.0.0.1:5000 ,页面支持录音、上传音频、生成纪要、导出、配置管理与一键「测试配置」。
四、方式二:MCP 协议(供 workbuddy 等调用)
标准 stdio 传输(本地 MCP 客户端最常用):
python mcp_server.pyHTTP 传输(远程调用,可选):
python mcp_server.py --http 127.0.0.1:8000workbuddy 集成示例
在 MCP 客户端配置中新增一个 stdio server,命令指向:
命令: D:\Python313\python.exe
参数: ["D:\\...\\mcp_server.py"]暴露的工具:
工具 | 功能 |
| 转写本地音频文件为中文文本 |
| 根据转写文本生成需求调研纪要(Markdown,含 Mermaid 流程图) |
| 导出纪要文件 md/html/pdf/docx |
| Whisper 模型管理 |
| 读取 / 修改全局配置 |
| 测试模型路径与 AI 后端连通性 |
五、方式三:命令行
# 转写音频
python cli.py transcribe --audio D:/rec.mp3 --model 中文优化版
# 生成纪要
python cli.py generate --text "客户说..." --backend openai --api-key nvapi-xxx
python cli.py generate --input transcript.txt --backend ollama
# 导出
python cli.py export --input minutes.md --fmt pdf
# 模型管理
python cli.py models list
python cli.py models add --name 中文优化版 --path D:/models/whisper-small-zh --desc "本地中文优化模型"
python cli.py models update --name 中文优化版 --path D:/models/whisper-small-zh
python cli.py models delete --name 中文优化版
# 配置管理
python cli.py config get
python cli.py config set --backend ollama --whisper-model small
# 测试配置
python cli.py test-config --backend openai所有命令输出 JSON,便于程序解析。
六、导出目录
生成的纪要文件默认保存在 exports/ 目录;录音文件在 uploads/;转写模型优先使用项目内 models/ 目录。
七、常见问题
转写很慢/无 GPU:改用内置
small模型;确认 ctranslate2 CUDA 库正常。OpenAI 接口 401:检查
.env的API_KEY与OPENAI_BASE_URL,用python cli.py test-config验证。Ollama 连接失败:先启动本地
ollama serve,并确认已ollama pull qwen2:7b(或改OLLAMA_MODEL)。
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- AlicenseAqualityAmaintenanceMCP server for audio transcription using local faster-whisper or OpenAI Whisper API, enabling multilingual transcription with optional GPT post-processing.Last updated3MIT
- Alicense-qualityDmaintenanceMCP server for audio transcription with speaker diarization. Transcribes MP3/WAV files using Faster-Whisper and pyannote.audio, outputs markdown with speaker labels, timestamps, summaries, and action items.Last updated1MIT
- Alicense-qualityAmaintenanceMCP server that enables audio transcription from files (wav, mp4, mp3, flac) or microphone recording, with dynamic tool selection and enterprise-grade security.Last updated2MIT
- Flicense-qualityCmaintenanceMCP server for whisper-based transcription and translation, supporting local stdio and remote HTTP transports with file workflow safety.Last updated
Related MCP Connectors
MCP server for generating rough-draft project plans from natural-language prompts.
MCP server for AI dialogue using various LLM models via AceDataCloud
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to interact with yo…
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/iaynidn-sudo/ai-translation-tool'
If you have feedback or need assistance with the MCP directory API, please join our Discord server