Space Integrations - Pulze.ai
Overview
Integrations extend your space capabilities by connecting external tools and services. Connect email, communication platforms, project management tools, and custom services to enhance your AI workflows.
Available Integrations
Gmail
Access and search your email conversations
Slack
Connect team communication channels
Access professional network data
MCP Tools
Custom integrations via Model Context Protocol
Gmail Integration
Connect your Gmail account to enable email-based AI assistance:
Setup
- Navigate to Integrations
Open your space settings and go to the Integrations section - Select Gmail
Click “Connect Gmail” button - Authorize Access
Sign in to your Google account and grant permissions - Configure Settings
- Select which labels to sync
- Set sync frequency
- Configure privacy settings
Capabilities
- Email Search: Find emails within conversations
- Email Context: Provide email threads as context to AI
- Smart Summaries: Generate email summaries
- Response Drafting: AI-assisted email composition
- Thread Analysis: Understand email conversation flow
Privacy & Security
- Read-only access by default
- Selective label access
- No automatic email sending
- Organization admin approval required
- Audit trail maintained
Slack Integration
Connect Slack to bring team communications into your AI workflow:
Setup
- Access Integrations
Navigate to space settings → Integrations - Connect Slack
Click “Connect Slack” button - Authorize Workspace
Approve Pulze app in your Slack workspace - Select Channels
Choose which channels to make accessible
Features
- Channel Search: Search across connected Slack channels
- Message Context: Reference Slack discussions in AI conversations
- Team Insights: Analyze team communication patterns
- Notification Integration: Get Pulze notifications in Slack
- Bot Commands: Interact with Pulze from Slack
LinkedIn Integration
Access LinkedIn data for research and outreach:
Setup
- Navigate to Integrations
Go to Integrations in space settings - Connect LinkedIn
Click “Connect LinkedIn” button - Authorize
Sign in and grant necessary permissions - Configure Access
Set privacy and data access preferences
Use Cases
- Profile Research: Analyze LinkedIn profiles
- Network Analysis: Understand connection patterns
- Content Research: Access shared content and posts
- Outreach Assistance: Draft personalized messages
- Company Intelligence: Research organizations
LinkedIn integration is organization-managed and requires admin approval.
MCP (Model Context Protocol) Tools
Connect custom tools and services using the Model Context Protocol:
What is MCP?
MCP is a standardized protocol for connecting external tools and data sources to AI models. It enables:
- Custom Tool Integration: Connect proprietary systems
- API Wrappers: Expose APIs as AI-accessible tools
- Data Sources: Connect databases and data stores
- Service Integration: Link external services
Setting Up MCP Tools
- Configure MCP Server
Set up your MCP server with custom tools - Register with Pulze
Add MCP server configuration to your space - Test Connection
Verify tools are accessible - Enable for Assistants
Allow assistants to use MCP tools
Example MCP Tools
- Database Queries: Direct database access for AI
- Internal APIs: Company-specific API tools
- File Systems: Access to document repositories
- Custom Services: Proprietary business tools
- Third-party APIs: Weather, maps, stock data, etc.
Integration Management
Viewing Integrations
The Integrations panel shows:
- Connected Services: Active integrations
- Status: Connection health
- Last Sync: When data was last updated
- Usage: How often the integration is used
- Settings: Configuration options
Disconnecting Integrations
To remove an integration:
- Navigate to the Integrations section
- Find the integration to remove
- Click “Disconnect”
- Confirm the action
Disconnecting an integration will remove access to that service’s data from all conversations and assistants in the space.
Reconnecting
If an integration fails:
- Check connection status
- Click “Reconnect”
- Reauthorize if needed
- Verify connectivity
Integration Permissions
Different integrations require different permission levels:
| Integration | Required Permission | Admin Approval |
|---|---|---|
| Gmail | Read emails | Yes |
| Slack | Read messages | Yes |
| Profile access | Yes | |
| MCP Tools | Varies by tool | Yes |
Integration Security
Data Privacy
- Selective Access: Choose what data to share
- Encrypted Transfer: All data encrypted in transit
- Access Logs: Track integration usage
- Revocable: Disconnect anytime
- Audit Trail: Complete activity history
Best Practices
- Only connect necessary integrations
- Review integration permissions regularly
- Monitor integration usage
- Disconnect unused integrations
- Follow organization security policies
Using Integrations in Conversations
Once connected, integrations enable new capabilities:
Automatic Tool Use
AI assistants can automatically:
- Search Gmail for relevant emails
- Find Slack messages for context
- Look up LinkedIn profiles
- Query MCP tools
- Combine multiple data sources
Manual Tool Invocation
You can explicitly request:
- “Search my Gmail for emails from John about the project”
- “Find Slack messages in #engineering about the bug”
- “Look up this person’s LinkedIn profile”
- “Query our database for customer data”
Integration Context
Integrations provide context by:
- Adding relevant data to conversations
- Enriching AI responses with real information
- Grounding outputs in actual data
- Reducing hallucinations
Integration Limits
Be aware of these limitations:
Rate Limits
- API calls per hour/day
- Search query limits
- Data retrieval caps
- Sync frequency restrictions
Data Limits
- Historical data range
- Number of results returned
- File size restrictions
- Storage quotas
Feature Limits
- Available endpoints
- Supported data types
- Query complexity
- Concurrent requests
Troubleshooting Integrations
Common Issues
Connection Failed
Solutions:- Verify credentials
- Check network connectivity
- Ensure correct permissions
- Try reconnecting
- Contact support if persists
Data Not Syncing
Solutions:- Check sync settings
- Verify service status
- Review integration logs
- Manually trigger sync
- Check for rate limiting
Permission Denied
Solutions:- Request organization admin approval
- Verify your user permissions
- Check integration scope
- Reauthorize connection
- Review access requirements
Slow Performance
Solutions:- Reduce sync frequency
- Limit data scope
- Optimize queries
- Check service health
- Consider caching options
Integration Analytics
Track integration usage:
- Call Volume: API requests per integration
- Data Transferred: Amount of data synced
- Success Rate: Successful vs failed requests
- Response Time: Integration latency
- Most Used: Popular integrations
Custom Integration Development
Build your own integrations:
MCP Server Development
- Design Tools: Define tool capabilities
- Implement Server: Build MCP-compliant server
- Test Locally: Verify functionality
- Deploy: Host your MCP server
- Register: Add to Pulze spaces