Skip to main content

Gmail

📧 Gmail Connector

Send, read, analyze, and automate workflows using your Gmail inbox.

The Gmail Connector lets your workflows interact with Gmail — allowing you to fetch incoming messages, filter them, analyze content using AI, send outbound emails, or forward information to other systems.

Use it to fully automate inbox-driven workflows, email assistants, reporting pipelines, and customer support flows.


🧠 How It Works

Gmail Connector Node

The Gmail Connector becomes a block in your workflow. When the workflow runs:

  1. The node executes Gmail actions (read, send, analyze).
  2. Filters and settings are applied (date range, unread, from, subject).
  3. The Gmail API is called securely through OAuth.
  4. The output data becomes available for the next nodes (AI, Slack, DB, etc.).

This makes it possible to chain Gmail with AI summarizers, spreadsheets, Slack notifications, or custom business logic.


⚙️ Configuration UI Guide

1️⃣ Credential Selection

Choose which Gmail OAuth connection will be used.

You may:

  • Select an existing credential
  • Search by provider
  • Search by workspace
  • Create a new credential directly from this modal

2️⃣ Gmail Account Settings

User Email

Auto-filled based on the connected Gmail account.

Max Emails (per run)

Maximum number of messages to fetch on each execution.

Default: 10

Days Back (search window)

How far back Gmail should search for messages.

Unread Only

Enable to fetch only unread messages.

Has Attachment

Enable to fetch only messages containing attachments.

These options make it easy to limit results to only relevant or recent emails.


3️⃣ Operations

Choose what the Gmail node should do when executed:

Read

Fetch email messages and return message metadata, content, and attachments.

Send

Send emails to one or more recipients (useful for notifications and automated reports).

Analyze

Apply AI-based analysis to message content (e.g., sentiment, metadata extraction, classification).

You can enable multiple operations at once.


4️⃣ Filters (for Read Operation)

Filters let you narrow which emails are fetched:

From (email)

Fetch only messages from a specific sender.

Subject contains

Match emails whose subject includes the given keyword/phrase.

Advanced query

Runs a raw Gmail search query (same syntax used in Gmail’s search bar).

Examples:

label:work from:billing@company.com has:attachment after:2025/01/01 subject:(invoice OR report)


5️⃣ Default Recipients (Send Operation)

When Send is enabled, you can predefine recipients for automated messages.

This is useful for:

  • Daily report emails
  • Notification workflows
  • Escalation emails
  • Auto-generated summaries

You can add multiple default recipients.


🚀 What You Can Do With the Gmail Connector

✨ Read Emails

Fetch unread messages, emails with attachments, or messages matching filters.

Use for:

  • Parsing OTPs, invoices, or customer requests
  • Daily summary workflows
  • AI-based content analysis

📝 Send Emails

Send auto-generated messages, AI-written responses, or workflow notifications.


🧠 Analyze Emails

AI-powered analysis of message text.

Use for:

  • Intent extraction
  • Sentiment detection
  • Auto-categorization
  • Extracting structured fields

📎 Read Attachments

Retrieve and process attachments like PDFs, CSVs, or images.


🛠️ Technical Operations

OperationFunction NameDescriptionRequired Parameters
Read Emailsgmail_read_emailsFetches emails using filters and search window.limit (optional)
Send Emailgmail_send_emailSends an email from the connected Gmail account.to, subject, body
Analyze Emailgmail_analyze_emailRuns AI analysis on message content.message_id
Read Attachmentsgmail_read_attachmentsDownloads attachments from an email.message_id

🔧 Authentication & Permissions

The Gmail Connector uses OAuth2 and requests the following scopes:

  • Read emails
  • Read metadata
  • Send emails
  • Read attachments

You may revoke access anytime from Google Account → Security.


🧩 Workflow Example: Fetch Recent Emails

{
"type": "gmailConnectorNode",
"config": {
"operations": ["read"],
"limit": 5,
"days_back": 7,
"unread_only": true
}
}