Compose reply draft in Gmail with OpenAI assistant

Open on n8n.io

$20/month : Unlimited workflows

2500 executions/month

Try free

THE #1 IN WEB SCRAPING

Scrape any website without limits

Try free

HOSTINGER

Early Deal
DISCOUNT 20%

Self-hosted n8n

Unlimited workflows - from $4.99/mo

Try free

#1 hub for scraping, AI & automation

6000+ actors - $5 credits/mo

Try free

Important notice

This workflow is provided as-is. Please review and test before using in production.

1. Workflow Overview

This workflow uses OpenAI Assistant to compose draft replies for labeled email messages. It automatically connects the drafts to Gmail threads. You can add knowledge base to your OpenAI Assistan...

Best for

Ticket Management automation workflows
Multimodal AI automation workflows
advanced n8n builders looking for reusable templates

Tools used

n8n-nodes-base.stickynote, n8n-nodes-base.gmail, n8n-nodes-base.httprequest, n8n-nodes-base.code, n8n-nodes-base.set, n8n-nodes-base.markdown, @n8n/n8n-nodes-langchain.openai, n8n-nodes-base.limit

Source and attribution

This workflow is cataloged by N8N Workflows and links back to its original n8n.io source page by Oskar.

Original n8n.io source

1.1 Workflow description

Title: Compose reply draft in Gmail with OpenAI assistant
Workflow name: Compose reply draft in Gmail with OpenAI assistant

This workflow uses OpenAI Assistant to compose draft replies for labeled email messages. It automatically connects the drafts to Gmail threads.

💡 You can add knowledge base to your OpenAI Assistant and make your reply drafts very customized (e.g. compose response with product information in response to inquiry from customer).

🎬 See this workflow in action in my YouTube video about automating Gmail.

How it works?

The workflow is triggered at regular intervals (default: every 1 minute – you can change this value) to check for messages with a specific label (e.g., "AI").

The content of the retrieved email message is then forwarded to the OpenAI Assistant node, and a reply draft is generated. Next, the response from the Assistant is converted to HTML, and a raw message in RFC standard is composed.

💡 You can learn more about composing drafts with the Gmail API in the official Google documentation.

The raw email message (reply draft) is encoded and attached to the original thread ID. Finally, the trigger label (in this case: "AI") is removed to prevent the workflow from looping.

Set up steps

Set credentials for Gmail and OpenAI.
Add new label in Gmail account for messages that should be handled by the workflow (e.g. name it "AI").
Select this label in the first and last Gmail nodes in workflow.
Create and configure your OpenAI Assistant. Select your assistant in "OpenAI Assistant" node.
Optionally: change trigger interval (by default interval is 1 minute).

If you like this workflow, please subscribe to my YouTube channel and/or my newsletter.

1.2 Logical Blocks

This catalog entry is organized from the workflow JSON. The node-level section below shows the executable blocks available for review before importing the template.

2. Block-by-Block Analysis

Block 1 - Sticky Note

Type / Role: n8n-nodes-base.stickyNote - stickyNote
Config choices: Version 1

Block 2 - Sticky Note2

Type / Role: n8n-nodes-base.stickyNote - stickyNote
Config choices: Version 1

Block 3 - Sticky Note1

Type / Role: n8n-nodes-base.stickyNote - stickyNote
Config choices: Version 1

Block 4 - Sticky Note4

Type / Role: n8n-nodes-base.stickyNote - stickyNote
Config choices: Version 1

Block 5 - Sticky Note5

Type / Role: n8n-nodes-base.stickyNote - stickyNote
Config choices: Version 1

Block 6 - Sticky Note7

Type / Role: n8n-nodes-base.stickyNote - stickyNote
Config choices: Version 1

Block 7 - Sticky Note8

Type / Role: n8n-nodes-base.stickyNote - stickyNote
Config choices: Version 1

Block 8 - Sticky Note9

Type / Role: n8n-nodes-base.stickyNote - stickyNote
Config choices: Version 1

Block 9 - Sticky Note10

Type / Role: n8n-nodes-base.stickyNote - stickyNote
Config choices: Version 1

Block 10 - Sticky Note11

Type / Role: n8n-nodes-base.stickyNote - stickyNote
Config choices: Version 1

Block 11 - Remove AI label from email

Type / Role: n8n-nodes-base.gmail - gmail
Config choices: Version 2.1

Block 12 - Add email draft to thread

Type / Role: n8n-nodes-base.httpRequest - httpRequest
Config choices: Version 4.2

Block 13 - Convert raw to base64

Type / Role: n8n-nodes-base.code - code
Config choices: Version 2

Block 14 - Build email raw

Type / Role: n8n-nodes-base.set - set
Config choices: Version 3.4

Block 15 - Convert response to HTML

Type / Role: n8n-nodes-base.markdown - markdown
Config choices: Version 1

Block 16 - Map fields for further processing

Type / Role: n8n-nodes-base.set - set
Config choices: Version 3.4

Block 17 - Ask OpenAI Assistant

Type / Role: @n8n/n8n-nodes-langchain.openAi - openAi
Config choices: Version 1.8

Block 18 - Get single message content

Type / Role: n8n-nodes-base.gmail - gmail
Config choices: Version 2.1

Block 19 - Return last message in thread

Type / Role: n8n-nodes-base.limit - limit
Config choices: Version 1

Block 20 - Get thread messages

Type / Role: n8n-nodes-base.gmail - gmail
Config choices: Version 2.1

Block 21 - Loop over threads

Type / Role: n8n-nodes-base.splitInBatches - splitInBatches
Config choices: Version 3

Block 22 - Get threads with specific labels

Type / Role: n8n-nodes-base.gmail - gmail
Config choices: Version 2.1

Block 23 - Schedule trigger (1 min)

Type / Role: n8n-nodes-base.scheduleTrigger - scheduleTrigger
Config choices: Version 1.2

3. Summary Table

Workflow	Compose reply draft in Gmail with OpenAI assistant
Complexity	advanced
Nodes	23
Categories	Ticket Management, Multimodal AI
Author	Oskar
Published	27 Mar 2024

4. Reproducing the Workflow from Scratch

1. Download the workflow JSON

Use the JSON export at /data/workflows/2198/2198.json as the source template for this automation.
2. Import the template into n8n

Open n8n, import the downloaded JSON, and review each node before activating the workflow.
3. Configure credentials and variables

Replace placeholder credentials, API keys, webhook URLs, account IDs, and environment-specific values with your own settings.
4. Test with sample data

Run the workflow manually or in a staging workspace, inspect node output, and confirm downstream systems receive the expected data.
5. Activate and monitor

Enable the workflow only after testing, then monitor executions, errors, and rate limits during the first production runs.

5. General Notes & Resources

Review imported nodes carefully before activation. This catalog entry is intended to help you inspect the workflow structure, understand required services, and find related templates faster.

Node names, credentials, schedules, webhook paths, and external service limits may need adjustment for your workspace.

Download workflow JSON Original n8n.io source Ticket Management workflows Multimodal AI workflows

Frequently asked questions

What does Compose reply draft in Gmail with OpenAI assistant do?

This workflow uses OpenAI Assistant to compose draft replies for labeled email messages. It automatically connects the drafts to Gmail threads. You can add knowledge base to your OpenAI Assistan...

What do I need before importing this workflow?

Review the workflow JSON, configure any required credentials in n8n, and test the automation in a safe workspace before using it in production.

Can I customize this workflow?

Yes. Use the block-by-block analysis and the downloadable JSON to inspect each node, then adjust credentials, prompts, schedules, filters, or destinations for your Ticket Management, Multimodal AI use case.

Oskar

4 workflows

Nodes

n8n-nodes-base.stickynote n8n-nodes-base.gmail n8n-nodes-base.httprequest n8n-nodes-base.code n8n-nodes-base.set n8n-nodes-base.markdown @n8n/n8n-nodes-langchain.openai n8n-nodes-base.limit

Complexity

advanced

Published 27 Mar 2024

Likes 0

View on n8n.io Download Workflow

Install path: /data/workflows/2198/2198.json

Share Your Workflow

Have a useful automation to share? Publish it and help the community.

Submit Your Template How to Submit

Related Workflows

Track post-therapy wellness with GPT-4o-mini, Telegram, Slack and Sheets

## Quick Overview This workflow uses Google Sheets, OpenAI, Telegram, and Slack to send post-therapy check-in messages, analyze patient replies for sentiment and risk, alert therapists to at-risk responses, and log non-urgent replies for ongoing wellness tracking. ## How it works 1. Runs every hour on a schedule trigger. 2. Reads session records from Google Sheets and keeps only sessions that have a session date and have not been marked as check-in sent. 3. Uses OpenAI (gpt-4o-mini) to generate a personalized, non-clinical Telegram check-in message from the patient’s session details. 4. Sends the check-in to the patient via Telegram and updates the Google Sheets session row to store the message and mark the check-in as sent. 5. Triggers on incoming Telegram messages, looks up the patient in a Google Sheets master list by Telegram ID, and selects the most relevant recent session to provide context. 6. Sends the patient reply and session context to OpenAI for JSON-formatted sentiment classification and parses the result into structured fields. 7. If the sentiment is classified as at-risk, posts an urgent alert to a Slack channel; otherwise, appends the reply and sentiment details to a Google Sheets wellness log. ## Setup 1. Connect Google Sheets OAuth credentials and update the spreadsheet document ID and sheet names for Sessions, Patients_Master, and Wellness_Log. 2. Create and connect a Telegram bot credential, set up the Telegram trigger webhook, and ensure patient Telegram chat IDs are stored in Google Sheets. 3. Add an OpenAI API key credential and confirm the selected model (gpt-4o-mini) meets your requirements. 4. Connect Slack OAuth credentials and replace the placeholder Slack channel ID used for therapist alerts, and confirm the error-alert channel is set correctly. 5. Verify your Google Sheets columns match the workflow’s expected fields (for example session_id, checkin_sent, checkin_message_sent, therapist_slack_id) before activating.

View

Detect and route customer escalations with Claude, Slack and Linear

## Quick Overview This workflow receives support messages via a webhook, normalizes payloads from Gmail, Zendesk, Intercom, and HelpScout, and uses Anthropic Claude to detect escalation signals and severity, then posts alerts to Slack and creates Linear tickets for actionable cases. ## How it works 1. Receives an inbound support message via a POST webhook. 2. Normalizes the incoming payload (Gmail, Zendesk, Intercom, HelpScout, or a raw test payload) into a consistent schema with customer details, subject, body text, and a source URL. 3. Sends the normalized message to Anthropic Claude and parses the response into structured JSON containing escalation flags, severity, trigger phrases, a summary, and a suggested reply opening. 4. Stops processing silently when the message is not an escalation (severity is none). 5. Builds a formatted Slack alert and Linear issue payload, including SLA guidance, detected signals, and links back to the original conversation. 6. Posts the escalation alert to a configured Slack channel. 7. Creates a Linear ticket for critical, high, and medium severity escalations and sends a Slack DM to a manager for critical and high cases. ## Setup 1. Add an Anthropic API credential for the Anthropic Chat Model node. 2. Add a Slack OAuth2 credential, set the target channel (for example, #escalations), and set the manager’s Slack user ID for direct messages. 3. Add a Linear OAuth2 credential and replace the placeholder Linear team ID used when creating tickets. 4. Copy the webhook URL from the webhook trigger and configure Gmail/Zendesk/Intercom/HelpScout (or your custom source) to POST message payloads to it in one of the supported formats.

View

Triage construction incidents with GPT-4o, Slack, Gmail and Google Sheets

## Quick Overview This workflow ingests construction incident webhooks, normalizes the payload, and uses OpenAI GPT-4o-mini to classify each incident by type and urgency, then routes alerts to Slack and Gmail and logs incidents to Google Sheets, with a daily Slack digest generated from the sheet. ## How it works 1. Receives an incoming incident via a POST webhook. 2. Normalizes the incoming Procore/Fieldwire-style payload into a consistent incident object (ID, title, description, location, reporter, assignee, and timestamps). 3. Sends the incident details to OpenAI (GPT-4o-mini) to classify issue type and urgency and generate a short summary, recommended action, and tags. 4. Routes incidents by urgency, sending Critical incidents to Slack with an @channel mention and emailing details to a project manager via Gmail. 5. Sends High-urgency incidents to Slack as a formatted notice. 6. Appends every incident to a Google Sheets log with status set to open. 7. Runs daily at 18:00 server time, reads the incident log from Google Sheets, builds an urgency-grouped digest, and posts it to a Slack channel. ## Setup 1. Add credentials for OpenAI, Slack (OAuth2), Gmail (OAuth2), and Google Sheets (OAuth2). 2. Copy the webhook URL from the Webhook trigger and register it in your source system (for example, Procore or Fieldwire) as the incident-created endpoint. 3. Replace the placeholder Slack channel IDs for Critical, High, Digest, and Error alerts. 4. Update the Gmail recipient address ([email protected]) to your real project manager email. 5. Set the Google Sheets spreadsheet (document ID) and ensure the target sheet exists with columns that match the incident log fields used by the workflow. 6. Confirm the schedule trigger time (cron: 0 18 * * *) matches your desired timezone/server time for the daily digest.

View

Need Custom Automation?

Get help designing a custom n8n workflow that connects your stack and fits your process.

Compose reply draft in Gmail with OpenAI assistant

Workflow preview

Important notice

1. Workflow Overview

Best for

Tools used

Source and attribution

1.1 Workflow description

How it works?

Set up steps

1.2 Logical Blocks

2. Block-by-Block Analysis

Block 1 - Sticky Note

Block 2 - Sticky Note2

Block 3 - Sticky Note1

Block 4 - Sticky Note4

Block 5 - Sticky Note5

Block 6 - Sticky Note7

Block 7 - Sticky Note8

Block 8 - Sticky Note9

Block 9 - Sticky Note10

Block 10 - Sticky Note11

Block 11 - Remove AI label from email

Block 12 - Add email draft to thread

Block 13 - Convert raw to base64

Block 14 - Build email raw

Block 15 - Convert response to HTML

Block 16 - Map fields for further processing

Block 17 - Ask OpenAI Assistant

Block 18 - Get single message content

Block 19 - Return last message in thread

Block 20 - Get thread messages

Block 21 - Loop over threads

Block 22 - Get threads with specific labels

Block 23 - Schedule trigger (1 min)

3. Summary Table

4. Reproducing the Workflow from Scratch

1. Download the workflow JSON

2. Import the template into n8n

3. Configure credentials and variables

4. Test with sample data

5. Activate and monitor

5. General Notes & Resources

Frequently asked questions