公告

Sonix MCP 服务器简介:将您的对话记录导入 Claude、Cursor 和 Codex

Sonix has long had the ability to use AI to interact with your transcripts, with one-click summaries, sentiment analysis, custom prompts, and more. For a lot of folks, that’s the quickest path from recording to insight, and Sonix makes it simple. But we know plenty of you spend your day inside another AI assistant: Claude, ChatGPT, Cursor, Codex… the list goes on. Wouldn’t it be nice to pull Sonix transcripts directly into your AI assistant workflow?

Well, now you can.

Sonix now speaks MCP

Sonix runs an MCP server, which is a fancy way of saying your favorite AI assistant can now talk to your Sonix account directly. You connect Sonix to a tool like Claude or Cursor once, and from then on your assistant can work with your media library on your behalf. No more copy-paste, glue scripts, or API keys to manage.

MCP (Model Context Protocol) is an open standard backed by Anthropic, OpenAI, and the broader AI tooling ecosystem. Connect Sonix to a single MCP-compatible client today and the same connection works in every new MCP client that ships tomorrow.

What you can ask your AI to do

Today the Sonix MCP server is read-only, which means your AI assistant can:

  • Browse your media library and find specific recordings
  • Pull a transcript into your conversation for summarization, Q&A, sentiment analysis, or any other LLM-powered analysis you’d normally use it for
  • Generate clean text transcripts, SRT or VTT subtitles, or structured JSON exports without ever leaving your chat
  • Check your Sonix plan and remaining credit

A few examples of what that looks like in practice:

“Find my interview with Dr. Patel from last week and pull out the three biggest takeaways.”

“Generate SRT subtitles for the most recent podcast episode.”

“Summarize all of my customer 面试 from this month into a single bulleted list.”

Which AI tools work with it?

Anything that speaks MCP. We’ve tested setup with Claude Code, Claude Desktop, Cursor, Codex, Windsurf, and VS Code, and our docs include copy-paste configuration snippets for each. When new MCP clients launch, if they speak the protocol, they’ll work with Sonix right out of the box.

Try it out

MCP access is included with every paid Sonix plan, at no extra cost. The setup takes about a minute:

  1. Point your AI tool at https://api.sonix.ai/mcp
  2. Sonix opens in your browser. Sign in and click Allow.
  3. Ask your assistant to do something with your media.

That’s the whole flow.

Head over to our MCP page for the bigger picture, or jump straight to the MCP section of our developer docs if you want to get your hands on it right now.

We’re excited to see what you build on top of it. As always, 给我们发个消息 if you’ve got feedback, run into issues, or want to tell us what you’d love to see next.

斯蒂芬·霍普金斯

最近的帖子

2026年每位内容创作者都应了解的22项播客文字稿增长统计数据

Comprehensive data compiled from extensive research on podcast transcription market growth, AI adoption, and content…

2周前

什么是自动转录?

Automated transcription is the process of converting spoken audio or video content into written text…

2周前

什么是发言人日记化?

Speaker diarization is an AI-powered process that automatically identifies and labels different speakers in audio…

2周前

原创音频。Yanny vs Laurel--这是我们的人工智能引擎的想法

我们通过Sonix AI引擎运行Yanny vs Laurel的音频文件,这里是...

2周前

什么是 Zoom 转录?

Zoom 转录是指将 Zoom 会议中的语音内容转换为书面文本的过程……

2周前

Sonix + Adobe Audition

Sonix has built the world's first AudioText Editor™ and it now works seamlessly with Adobe…

2周前

本网站使用 cookie。