Technology & AI

Build an AI Assembly Summary with Claude Code + MCP

Teams in all companies lose meeting notes and action items after discussions. This guide builds a permanent fix: Summary of the AI ​​Assembly and Task Scheduler using Claude Code and MCP. It processes the transcripts into organized summaries with tasks, decisions, and calendar reminders, connects to Google Calendar and Gmail, and stores everything in SQLite.

MCP acts as a universal adapter between Claude and external systems, keeping the integration consistent. In this article, you will learn how to create and implement this integrated end-to-end workflow.

What is Model Context Protocol (MCP)?

The Model Context Protocol (MCP) enables Claude to interact with external systems through a schema-based interface that acts as an open standard. The MCP system provides Claude with a reliable way to access resources through a dedicated calling interface that works as an open standard.

The project uses MCP as its main framework. Claude accesses Google Calendar and Gmail functions through MCP tools that provide access to those services. The system maintains clean agent logic that developers can easily inspect and modify. You need to change the MCP server if you want to switch from Gmail to Outlook, but you don’t need to change the agent code.

The concept of MCPWhat it doesFor example in this project
tools/listDiscover the tools available at runtimeAgent receives create_event, send_email
tools/telephoneRequest a tool with structured JSON inputClaude calls creative_event with title and time
inputSchemaValidates all inputs before executionPrevents incorrect calendar entries
tool_resultReturns structured output back to ClaudeReturns the ID of the event after the calendar is created
isErrorSigns of failure of the instrument level with kindnessReturns isError: true at API rate limit

Agent Architecture

The Meeting Summary and Scheduling Agent will follow a multi-stage pipeline. Each section is isolated from the area, can fail, retry, or skip independently. This is what makes it a production grade and not just for demo purposes.

The stageElementResponsibility
ImportText LoaderUpload the .txt/.vtt/.json transcript, adjust the text to normal
AnalyzeClaude (LLM)Extract summary, decisions, owners, deadlines at the prompt
ConfirmSchema ValidatorVerify the output matches the expected JSON structure
PersistSQLite storeWrite meeting record, action items, decisions in DB
ShippingMCP Tool WrapperCall Calendar MCP to create events, Gmail MCP to send report
ConfirmAuditor of effectVerify successful tool calls; the line tries again with failure

Getting started

The program uses Claude Code as its AI agent that operates through a user interface. MCP servers enable Claude Code to interact with external systems by acting as its own control mechanisms. You connect Google Calendar and Gmail as MCP servers and. After that, every session, Claude Code can call them just for you asking in plain English, no API wrappers, no async loops, no tool dispatch code.

The program enables Claude Code to access more than 300 external tools through its MCP connection. The program allows you to issue commands like “Create a Gmail draft and invite users to a feedback session” after which it handles all the remaining tasks. Follow the simple steps below and you’ll have an agent up and running in no time:

Step 1: Enter the Claude Code

npm install -g @anthropic-ai/claude-code
claude --version   # confirm install
claude         # launch your first session 

You will enter an interactive terminal session after completing the login process. This agent goal allows you to send messages that will be converted into Claude’s commands. The program provides access to all slash commands and MCP tools through this interface.

Step 2: Set the Project Folder

Create a dedicated folder for the project. The project uses a CLAUDE.md file that resides in the root of the project to serve as a permanent memory of the Claude Code and instruction base. This area serves as a place where you provide all the project information Claude needs to know for future reference.

mkdir meeting-agent && cd meeting-agent  

# Create the folder structure  
mkdir -p transcripts summaries .claude/commands  

# Open Claude Code in this project  
claude

Now ask the cloud model to create a CLAUDE.md file. Here is a warning you can use:

Create a CLAUDE.md file for this project. This project analyzes the conference documents. Reads .txt or .vtt files in /written folder, extracts action items by owners and due dates, creates Google Calendar events with MCP, sends Gmail follow-up emails with MCP, and writes a summary /abbreviations. The MCP servers are named ‘gcal’ and ‘gmail’. Always provide a deadline from context: defaults to 3 business days if not specified. Owners must match the names mentioned in the transcript.

Validation.json

The system performs its main functions in this area. Claude Code’s built-in mcp command allows users to register MCP servers through the use of a terminal. The process requires the execution of one command on each server.

3a. Connect Google Workspace MCP (Calendar + Gmail)

The google_workspace_mcp server offers two services with its single server that includes Calendar and Gmail functions. The process starts with setting up Google OAuth authentication in the Google Cloud Console before moving on to server registration.

Step 1: Add MCP server to Claude Code

claude mcp add --transport stdio google-workspace  
-- npx -y google-workspace-mcp  
--client-id YOUR_GOOGLE_CLIENT_ID  
--client-secret YOUR_GOOGLE_CLIENT_SECRET ] 

Step 2: Make sure it is registered correctly

claude mcp list

Expected output:

google-workspace stdio npx -y google-workspace-mcp
Gmail and Google Calendar connectors

3b. Verify with Google

The first time you use the MCP server requires you to authenticate with your Google account. Use the auth command within a Claude Code session to enable authentication.

The server’s authentication process will begin with a Code Claude that requires users to open a web browser. The user needs to login and provide the necessary permissions before the tokens are saved to their device. You only do this once.

3c. Verify Tools Are Available

Inside Claude’s Code, use the builtin /mcp command to see all the loaded tools:

# Inside Claude Code session  
/mcp  

# You should see something like:  
# google-workspace (connected)  
#   create_calendar_event  
#   update_calendar_event  
#   send_email  
#   create_draft  
#   list_calendar_events  
#   ... and more

Step 4: Create /process-meeting Slash Command

Slash commands in Claude Code markup files are stored in .claude/commands/. The file name becomes the command name. Content is a notification that Claude uses when you use it. You can use this method to convert a complex process that requires many steps into a single command.

# Create the slash command file  
touch .claude/commands/process-meeting.md

Now ask Claude Code to write a command for you, this is the correct way to do it:

Write a slash command file for .claude/commands/process-meeting.md. Command takes $ARGUMENTS as a text file name. The process should

  1. read the file from /transcripts/$COMMUNICATIONS
  2. remove all action items by owner (task) deadline (and priority)
  3. create a Google Calendar event for each action item using the MCP server
  4. send a Gmail to each owner summarizing their activities
  5. write a clean summary into /summaries. A detailed command needs to meet standard production requirements.
process-meeting.md file
Running the steps in sequence

Step 5: Using the agent

Place a sample transcript in /written folder and run the command. That’s the whole user experience.

# Copy your transcript in  
cp ~/Downloads/team-standup-march24.txt transcripts/  

# Open Claude Code  
claude  

# Run the slash command  
/process-meeting team-standup-march24.txt

Output:

Google workplace events are done through MCP
Reading transcripts and prioritizing actions

The conclusion

What he builds traditionally takes weeks: DB schema, pipes, retries, OAuth, APIs. With Claude Code and MCP, it was done in an afternoon using prompts and a few CLI commands.

The model is simple: you are the Premier, Claude Code is the developer, MCP servers are ready for integration. He explains the requirements with CLAUDE.md, slash commands, and commands. Claude delivers, handling real complexity and scaling with just new instructions.

Frequently Asked Questions

Q1. What is Model Context Protocol (MCP)?

IA. MCP is a common interface that allows Claude to connect to external tools like Gmail and Google Calendar without custom integration.

Q2. What does an AI assembly summary agent do?

A. Processes transcripts, issues action items, saves them, creates calendar events, and sends follow-up emails automatically.

Q3. What is the role of slash commands in Claude Code?

A. They turn multi-step workflows into single-command, automated tasks like summarizing meetings and triggering MCP tools.

Riya Bansal

Data Science Trainee at Analytics Vidhya
I currently work as a Data Science Trainer at Analytics Vidhya, where I focus on building data-driven solutions and applying AI/ML techniques to solve real-world business problems. My work allows me to explore advanced analytics, machine learning, and AI applications that empower organizations to make smarter, evidence-based decisions.
With a strong foundation in computer science, software development, and data analysis, I am passionate about using AI to create impactful, innovative solutions that bridge the gap between technology and business.
📩 You can also contact me at [email protected]

Sign in to continue reading and enjoy content curated by experts.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button