Connect Hedy with AI Assistants Using MCP
Access your Hedy sessions, highlights, and tasks directly from AI assistants using the Model Context Protocol (MCP).
What is the Hedy MCP Server?
The Hedy MCP server enables AI assistants to securely access your Hedy data. Once connected, you can ask your AI to analyze meeting transcripts, find highlights, review action items, and organize sessions—all through natural conversation.
MCP (Model Context Protocol) is an open standard that allows AI applications to connect to external data sources. The Hedy MCP server uses OAuth authentication, making it compatible with web-based AI platforms, desktop applications, and development tools.
How to Connect
Step 1: Add the Server
Open your AI client's MCP server settings
Look for "Add MCP Server", "Connect Remote Server", or "Add Integration"
Enter the server URL:
https://api.hedy.bot/mcp
Click "Add" or "Connect"
Step 2: Authorize Access
Your AI client will open a browser window
Sign in with your Hedy account
Review the requested permissions
Click "Allow" to authorize
Return to your AI client
That's it! No configuration files or API keys needed.
Compatible AI Clients
Works with any MCP-compatible client that supports remote servers:
Web-based AI platforms - Browser-accessible AI assistants
Desktop applications - Claude Desktop, Cursor, Windsurf, Zed
Development tools - VS Code with GitHub Copilot, Cline
AI agent platforms - Custom workflow builders
Available Tools
The Hedy MCP server provides 10 tools your AI assistant can use:
Session Tools
GetSessions
What it does: Lists your recent meeting sessions, ordered by start time (newest first)
Parameters:
limit
(optional) - Number of sessions to return (1-100, default varies by client)after
(optional) - Session ID to start pagination after
Example requests:
"Show me my last 10 sessions"
"List all sessions from this week"
"What meetings did I have recently?"
GetSessionDetails
What it does: Retrieves full details for a specific session including metadata, participants, and transcript
Parameters:
sessionId
(required) - The ID of the session
Example requests:
"Show me the details of session ABC123"
"What was discussed in my last team meeting?"
"Get the transcript for the product launch session"
GetSessionHighlights
What it does: Returns all highlights captured during a specific session
Parameters:
sessionId
(required) - The ID of the session
Example requests:
"What highlights did I save in session ABC123?"
"Show me the key moments from yesterday's standup"
"List all highlights from the budget meeting"
GetSessionToDos
What it does: Lists all to-do items generated during a specific session
Parameters:
sessionId
(required) - The ID of the sessionincludeCompleted
(optional) - If true, includes completed tasks (default: only open tasks)
Example requests:
"What action items came from session ABC123?"
"Show me todos from the client meeting, including completed ones"
"List pending tasks from yesterday's planning session"
Highlight Tools
GetHighlights
What it does: Lists your recent highlights across all sessions, sorted by newest first
Parameters:
limit
(optional) - Number of highlights to return (1-100)after
(optional) - Highlight ID to start pagination after
Example requests:
"Show me my last 20 highlights"
"What have I highlighted this week?"
"List recent key moments I've saved"
GetHighlightDetails
What it does: Retrieves full details for a specific highlight including transcript snippet, timestamp, and context
Parameters:
highlightId
(required) - The ID of the highlight
Example requests:
"Show me highlight XYZ789 in detail"
"What was the context around that budget highlight?"
"Get the full transcript for highlight XYZ789"
Task Tools
GetToDos
What it does: Returns all to-do items across all your sessions
Parameters:
includeCompleted
(optional) - If true, includes completed tasks (default: only open tasks)
Example requests:
"What are all my pending action items?"
"Show me all todos, including completed ones"
"List everything I need to do from my meetings"
Topic Tools
GetAllTopics
What it does: Lists all your meeting topics and folders
Parameters:
includeArchived
(optional) - If true, includes archived topics (default: only active topics)
Example requests:
"What topics do I have in Hedy?"
"List all my meeting categories"
"Show me all topics including archived ones"
GetTopicDetails
What it does: Retrieves details for a specific topic including description and statistics
Parameters:
topicId
(required) - The ID of the topic
Example requests:
"Show me details for the Client Meetings topic"
"How many sessions are in topic DEF456?"
"What's in my Team Standups topic?"
GetTopicSessions
What it does: Lists all sessions associated with a specific topic
Parameters:
topicId
(required) - The ID of the topic
Example requests:
"Show me all sessions in the Product Planning topic"
"List meetings tagged with topic DEF456"
"What sessions are in my Client Meetings folder?"
Example Workflows
Weekly Review
Ask your AI: "Show me my sessions from this week, list all pending todos, and highlight any important items"
What happens: AI uses GetSessions with date filtering, then GetToDos to find action items, and can drill into specific sessions for details.
Meeting Preparation
Ask your AI: "Find all sessions about the Q4 roadmap and summarize the key points"
What happens: AI searches through sessions, uses GetSessionDetails and GetSessionHighlights to gather information, then creates a summary.
Task Management
Ask your AI: "Show me all action items from client meetings this month"
What happens: AI uses GetAllTopics to find the Client Meetings topic, GetTopicSessions to list sessions, then GetSessionToDos for each to compile the list.
Topic Analysis
Ask your AI: "Analyze my Team Standup sessions and identify recurring themes"
What happens: AI finds the topic, retrieves all sessions in it, reads transcripts and highlights, then performs analysis.
Client-Specific Setup
Claude Desktop
Go to Settings → Developer → Edit Config
Add this configuration:
{ "mcpServers": { "hedy": { "url": "https://api.hedy.bot/mcp", "transport": "http" } } }
Restart Claude Desktop
Follow authorization prompts
Web-Based AI Clients
Find "Connect MCP Server" or "Add Integration"
Enter:
https://api.hedy.bot/mcp
Complete OAuth authorization in your browser
VS Code / Cursor / Other Tools
Open MCP settings
Add server URL:
https://api.hedy.bot/mcp
Select "HTTP with OAuth" as connection type
Authorize when prompted
Privacy and Security
OAuth 2.1 with PKCE - Secure industry-standard authentication
Read-only access - Cannot delete or modify your data
On-demand only - Data accessed only when you make requests
Troubleshooting
AI client doesn't have "Add MCP Server" option
Check if your client supports remote MCP servers
Update to the latest version
Look for MCP settings in preferences
Some clients only support local servers and need manual configuration
Authorization fails
Ensure you're signed in to Hedy
Verify your account is active
Clear browser cache and retry
Check internet connection
Try opening
https://api.hedy.bot/mcp/.well-known/oauth-protected-resource
to verify server is accessible
AI can't access my data
Confirm OAuth authorization completed
Verify you have sessions in your Hedy account
Check authorization status in AI client settings
Try disconnecting and reconnecting
Connection stopped working
Access token may have expired—reauthorize
If you changed Hedy password, reauthorize
Check
status.hedy.bot
for service issuesRemove and re-add the server
Client requires manual OAuth configuration
If your client doesn't support automatic discovery, use these settings:
Authorization URL:
https://api.hedy.bot/mcp/oauth/authorize
Token URL:
https://api.hedy.bot/mcp/oauth/token
Scopes:
mcp.sessions.read mcp.highlights.read mcp.todos.read mcp.topics.read
AI says a tool failed or returned an error
Check that the session/highlight/topic ID is valid
Verify you have permission to access that data
Ensure the requested data exists in your account
Try asking more specifically (e.g., include session titles instead of IDs)
Check for rate limiting—wait a moment and retry
Best Practices
For Better Results
Use clear session titles - Makes them easier for AI to find
Organize with topics - Group related sessions together
Create highlights during meetings - Helps AI identify key moments
Add descriptive todo notes - Makes action items more searchable
Be specific in requests - Include dates, topics, or keywords
When Asking Your AI
Specify time ranges - "Show sessions from last week" is clearer than "Show recent sessions"
Mention topics when relevant - "Client meeting sessions" vs "all sessions"
Ask for summaries - AI can synthesize data from multiple tools
Request specific formats - "Create a bullet list" or "Make a table"
Follow up with questions - AI can drill deeper into results
Pro tip: The AI can combine multiple tools in one request. Try asking complex questions like "Find all pending todos from client meetings this month and organize them by priority" to see the power of MCP integration.
Need Help?
Check your AI client's MCP documentation for client-specific instructions
Visit
modelcontextprotocol.io
for general MCP informationContact Hedy support with error messages and your AI client name