Use Speak AI with Cursor — Transcription & Analysis in Your IDE
Access your Speak AI transcripts, recordings, and analysis directly in Cursor. Search across meetings, pull insights, and manage your media library without leaving your editor.
Trusted by teams at
What you can do in Cursor
The Speak AI MCP server gives Cursor's AI assistant direct access to your transcripts, recordings, and analysis tools. Here is what that looks like in practice.
Search transcripts while coding
Find what the PM said about requirements, what customers reported in interviews, or specific feature requests across your team's recordings. Search by keyword, speaker, or topic without switching windows. Results come back right in your Cursor conversation.
Transcribe and analyze
Upload a recording from Cursor and get the transcript, key topics, sentiment, and action items returned directly in your conversation. No need to open a browser, upload a file somewhere else, and copy-paste results back. The entire workflow stays inside your editor.
Reference meeting context
Pull meeting summaries and decisions into your coding context. Let Cursor's AI reference real conversations when helping you build. When you need to know why a design decision was made or what the client actually asked for, the answer is one prompt away.
Manage your workspace
Create folders, organize recordings, schedule meeting bots, and export transcripts. All through Cursor's AI assistant. You get full workspace control without switching to the Speak AI dashboard for routine tasks.
Set up Speak AI in Cursor
Three steps from zero to searching your transcripts inside Cursor. The MCP server auto-detects your IDE and handles configuration.
Install the MCP server
Install the Speak AI MCP server globally from npm. Run the init command and it will auto-detect Cursor and write the configuration for you. No manual JSON editing needed.
Get your API key
Sign up at speakai.co if you do not have an account yet. Then go to Settings > API and copy your API key. The init command will prompt you for it during setup.
Start using it in Cursor
Open a Cursor chat and try commands like: "@speak transcribe this recording", "search my meetings for auth discussion", or "what did the client say about the deadline?" The AI assistant will use your Speak AI data to answer.
npm install -g @speakai/mcp-server
speakai-mcp init # auto-detects Cursor and configures
# Then in Cursor chat:
"Search my meetings for the auth discussion"
"Transcribe this recording and pull out action items"
"What did the client say about the deadline?"
Built for how developers actually work
Every team generates hours of meeting recordings. The Speak AI integration puts that context where you need it: inside your IDE.
Product development
Reference user interview transcripts while building features. When you are implementing a feature and need to know exactly what users said about it, pull the relevant transcript segments into your Cursor conversation. Build from real user feedback, not secondhand summaries in a ticket.
Code reviews
Pull meeting context about design decisions into review discussions. When a PR looks unusual and you need to understand why it was built that way, search for the architecture meeting where the approach was decided. Add that context directly to your review.
Sprint planning
Search across retrospective and planning meeting transcripts to find patterns. What keeps coming up in retros? What did the team agree to try differently? Get real data from past conversations instead of relying on memory.
Speak AI and Cursor: transcription and analysis inside your editor
Cursor is an AI-powered code editor built on top of VS Code. It is designed for developers who want to use large language models as part of their daily coding workflow. Cursor supports the Model Context Protocol (MCP), which lets external tools and data sources connect directly to the AI assistant built into the editor. That is where Speak AI comes in.
The Speak AI MCP server connects your entire media library, transcripts, and analysis tools to Cursor. Once configured, you can search transcripts, upload recordings, run NLP analysis, and manage your workspace through natural language prompts in the Cursor chat. There is no context switching between your IDE and a separate transcription dashboard. The data comes to you.
How MCP works in Cursor
MCP (Model Context Protocol) is an open standard that lets AI assistants access external tools and data. Cursor's MCP support means any MCP-compatible server can extend what the AI assistant can do. When you install the Speak AI MCP server, Cursor gains access to 81 tools across 12 categories: media management, AI chat, folders, recorder and survey, automations, clips, custom fields, webhooks, meeting assistant, embeds, text notes, and exports. The server handles authentication and API calls behind the scenes.
In practice, this means you can type a natural language request in the Cursor chat panel and the AI figures out which Speak AI tools to call. Ask "find all mentions of the pricing discussion from last week's meetings" and the MCP server searches your transcripts, filters by date, and returns the relevant segments. Ask "transcribe this file and give me the key topics" and it uploads, transcribes, runs NLP, and summarizes the results.
What Speak AI adds to your Cursor workflow
Developers sit in meetings, standups, design reviews, and client calls every day. Those conversations contain decisions, requirements, and context that matter when writing code. But that information usually lives in a separate app or gets lost in someone's notes. With Speak AI connected to Cursor, every recorded conversation becomes searchable context.
When you are implementing a feature, you can pull up what the product manager said about the requirements in the last sprint planning call. When reviewing a pull request, you can search for the architecture discussion where the team decided on the approach. When onboarding to a new codebase, you can ask for summaries of recent technical discussions to get up to speed faster. The integration supports transcription in over 70 languages with speaker identification, sentiment analysis, keyword extraction, and topic detection.
Setup and configuration
Installation takes about two minutes. Install the MCP server from npm with npm install -g @speakai/mcp-server, then run speakai-mcp init. The init command detects that you are running Cursor and writes the MCP configuration automatically. You will need a Speak AI API key, which you can get by signing up at speakai.co and going to Settings > API. The trial gives you full access to all 81 MCP tools, so you can evaluate the integration before committing to a paid plan.
The Speak AI MCP server is open source under the MIT license and available on GitHub. It also works with Claude Desktop, ChatGPT, Windsurf, VS Code, and any other MCP-compatible client. If your team uses multiple editors or AI tools, one installation covers all of them.
Beyond transcription
Speak AI is not just a transcription tool. The platform includes automated meeting bots that join your calls and record them, an embeddable recorder for surveys and interviews, NLP analytics that extract sentiment, themes, keywords, and named entities, and multi-model AI Chat that supports Claude, Gemini, and GPT. All of these capabilities are accessible through the MCP server in Cursor. You can schedule a meeting bot from your editor, analyze a batch of customer interviews, or ask questions across your entire transcript library. It is a full conversational data platform connected to the place where you spend most of your working hours.
Frequently asked questions
Common questions about using Speak AI with Cursor IDE.
How do I add Speak AI to Cursor?
Install the Speak AI MCP server globally with npm install -g @speakai/mcp-server, then run speakai-mcp init. The init command auto-detects Cursor and writes the configuration file for you. You will need a Speak AI API key, which you can get by signing up at speakai.co and navigating to Settings > API. The entire setup takes about two minutes.
What can I do with Speak AI in Cursor?
You get access to 81 tools that let you search transcripts, upload and transcribe recordings, run NLP analysis (sentiment, keywords, topics, entities), manage folders and recordings, schedule meeting bots, export data, and ask questions across your entire media library. Everything works through natural language prompts in the Cursor chat panel.
Does it slow down my IDE?
No. The MCP server runs as a lightweight background process and only activates when you make a request through the Cursor AI chat. It does not index files, watch directories, or consume resources when idle. API calls are processed by Speak AI's servers, not on your local machine.
Is it free?
The MCP server itself is free and open source under the MIT license. You need a Speak AI account to use it. Speak AI offers a free 7-day trial with full access to all 81 tools, so you can evaluate the integration before choosing a plan. After the trial, paid plans start at $15/month. See pricing details.
Can I use it alongside other MCP servers?
Yes. Cursor supports multiple MCP servers running simultaneously. Speak AI's MCP server will work alongside any other MCP servers you have configured, such as database tools, documentation servers, or other integrations. Each server operates independently and the AI assistant can use tools from all connected servers in a single conversation.
Bring your transcripts into Cursor
Every meeting, interview, and conversation your team records becomes searchable context inside your editor. Set up in two minutes, free to try.
Try Speak AI free
Create an account and get full access to all 81 MCP tools for 7 days. No credit card required. Install the MCP server, connect to Cursor, and start searching your transcripts immediately.
View the MCP server
Browse all 81 tools, see code examples, and read the full documentation. Open source on GitHub under the MIT license. Works with Cursor, Claude, ChatGPT, Windsurf, and VS Code.