> ## Documentation Index
> Fetch the complete documentation index at: https://docs.beam.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Integration Connectors
> Connect agents to external services with pre-built integration connectors for CRM, communication, productivity, and automation
Integration connectors enable your agents to interact with external platforms like Salesforce, Gmail, Airtable, and Slack. Configure parameters using variable fill methods to pass data between your agent and these services.
Learn how to add integrations to your workflows in [Selecting Tools](/02-building-agents/agent-fundamentals/tools-integrations/tools-integrations).
## Configuring Parameters
Each integration action expects specific input parameters. Every integration provides parameter description guidelines that explain the required format.
### Understanding Parameter Descriptions
Integration tools display parameter requirements with format guidelines directly in the configuration UI.
**Parameter Information:**
* **Name**: Field identifier (e.g., `site_name`, `search`, `folder_path`)
* **Description**: Format requirements and examples provided by the integration
* **Required/Optional**: Whether the parameter must be provided
* **Fill Method**: How the value will be populated
### Fill Method Selection
Choose the fill method based on your data source:
**Linked Variables** - Pass data from previous nodes:
```
${previous_node.output_field}
```
**AI Fill** - AI constructs value from available context using the parameter description
**Static** - Fixed values that never change
**User Fill** - Runtime input from user
**From Memory** - Retrieved from agent's memory
### AI Fill Configuration
When using AI Fill, the parameter description provided by the integration guides the AI on how to format the output.
**Example: Google Sheets Append Row**
The integration provides this description for the `data` parameter:
```
Array of comma-separated value strings for each new row,
e.g. ["Alice","Bob","100", "Carol","Dave","200"]
```
**Your AI Fill Configuration:**
```
Extract customer name, email, and order total from the conversation.
Format as array of strings following the example format.
```
The AI combines the integration's format requirement with your extraction instructions.
**Example: Salesforce Update Record**
The integration provides this description for the `fields` parameter:
```
Key-value pairs of fields to update, e.g.
{"Name":"New Name","Phone":"1234567890"}
```
**Your AI Fill Configuration:**
```
Extract customer information from the context.
Use exact Salesforce API field names (FirstName, LastName, Email, Phone).
Format as JSON object per the example.
```
**Example: SharePoint Search Folders**
The integration provides this description for the `search_query` parameter:
```
Search query string to filter folders.
Use SharePoint search syntax (e.g., "ContentType:Folder AND Title:*Report*")
```
**Your AI Fill Configuration:**
```
Extract the folder name or keywords from user's request.
Format as SharePoint search query targeting folders only.
If searching for "Q4 Reports", use: "ContentType:Folder AND Title:*Q4*Report*"
```
### Writing Effective AI Fill Descriptions
**Start with the integration's format requirement**, then add your extraction logic.
**Poor:**
```
"Get customer data"
```
**Good:**
```
"Array format: ['Name','Email','Phone'] per integration requirement.
Extract from conversation and format accordingly."
```
**Use the exact examples** provided by the integration.
**Poor:**
```
"JSON object with fields"
```
**Good:**
```
"JSON format: {'FirstName':'John','LastName':'Doe'} as shown in integration example.
Extract from customer context."
```
**Use exact API field names** as shown in the integration description.
**Poor:**
```
"Contact information"
```
**Good:**
```
"Use exact field names: FirstName, LastName, Email, Phone (not 'first name' or 'e-mail').
Salesforce is case-sensitive."
```
**Add fallback behavior** when data might be missing.
**Poor:**
```
"Extract items"
```
**Good:**
```
"If items found, format as ['Item1','Item2']. If no items, return empty array [].
Follow integration's array format."
```
### Linked Variables from Structured Outputs
Use Linked when previous nodes output clean, structured data that matches the integration's format.
**Example Workflow:**
```
Node 1: Extract Customer Info (Custom GPT)
├─ Structured Outputs:
│ ├─ customer_first_name: "John"
│ ├─ customer_last_name: "Doe"
│ ├─ customer_phone: "555-1234"
│ └─ salesforce_record_id: "003XX000004TmiI"
Node 2: Update Salesforce (Integration)
├─ Fields (Linked): {"FirstName":"${extract.customer_first_name}","LastName":"${extract.customer_last_name}","Phone":"${extract.customer_phone}"}
├─ Object Type (Static): "Contact"
└─ Record ID (Linked): ${extract.salesforce_record_id}
```
### Integration-Specific Format Patterns
Different integrations require specific data structures. Always check the parameter description.
**Arrays for List Operations:**
```
Google Sheets, Airtable: ["Item1", "Item2", "Item3"]
```
**JSON Objects for Record Operations:**
```
Salesforce, HubSpot: {"Field": "Value", "AnotherField": "Value"}
```
**Linked Record IDs:**
```
Airtable: ["recABC123"] // Must be array even for single link
Salesforce: "003XX000004TmiI" // String ID
```
**Search Queries:**
```
SharePoint: "ContentType:Folder AND Title:*keyword*"
Google Drive: "mimeType='application/vnd.google-apps.folder' and name contains 'Reports'"
```
## Integration Catalog
Beam provides 1500+ pre-built integrations organized by category.
Salesforce, HubSpot, Pipedrive, Zoho, Copper
Slack, Teams, Zoom, Discord, LinkedIn
Google Workspace, Microsoft 365, Notion, Calendly
Gmail, Outlook, SendGrid, Mailchimp
Google Sheets, Airtable, Excel, Smartsheet
Google Drive, Dropbox, Box, OneDrive, SharePoint
Mailchimp, ActiveCampaign, Klaviyo
Freshdesk, Zendesk, Jira, Asana
Shopify, WooCommerce, Stripe, PayPal
Google Analytics, Mixpanel, Segment
GitHub, GitLab, Bitbucket, Linear
OpenAI, Anthropic, Hugging Face
### Featured Integrations
CRM platform for accounts, contacts, leads, and opportunities. Automate record creation, updates, and searches.
Email automation for sending, searching, drafting, and managing messages with Google Workspace integration.
Spreadsheet operations for appending rows, updating cells, reading data, and managing worksheets.
Team messaging for notifications, channel creation, file uploads, and thread posting.
Database management for creating records, updating fields, querying data, and linking records.
Marketing and sales automation for contacts, deals, companies, and tickets.
Workspace management for creating pages, updating databases, and organizing content.
Payment processing for charges, subscriptions, customers, and refunds.
Customer support for ticket management, user updates, and issue tracking.
E-commerce operations for products, orders, customers, and inventory.
## Troubleshooting
**Issue:** Integration rejects parameter format
**Solutions:**
* Check integration's parameter description for exact format
* Follow the examples provided by the integration
* Verify data types match (string vs number vs boolean)
* Ensure arrays use correct structure (`["item"]` vs `"item"`)
* Review integration documentation for field naming
**Issue:** AI generates data in incorrect structure
**Solutions:**
* Reference the integration's format example in your description
* Be explicit: "Follow this exact format" with specific example
* Specify data types as shown in integration guidelines
* Add validation: "Must match example format exactly"
* Consider using Linked from structured output instead
**Issue:** Cannot authenticate or connection expired
**Solutions:**
* Verify credentials are current and valid
* Re-authenticate OAuth connections
* Confirm API key has required permissions
* Check OAuth scopes match integration requirements
**Issue:** Integration doesn't recognize field names
**Solutions:**
* Use exact API field names from integration description
* Check case sensitivity (e.g., `FirstName` not `firstname`)
* Verify field exists in integration's schema
* Review integration documentation for field mapping
## Next Steps
Learn how to add integrations to workflows
Understand variable fill methods in detail
Design outputs for clean integration inputs
Build custom API connections