klbr/n8nworkflows.xyz

Fork 0

mirror of https://github.com/khoaliber/n8nworkflows.xyz.git synced 2026-04-28 00:29:22 +00:00

Files

T

nusquama fcf7e8fb80 creation

2025-11-13 20:10:33 +01:00

16 KiB

Raw Blame History

Automatically Categorize and Label Gmail Emails with OpenAI GPT-4 Classification

https://n8nworkflows.xyz/workflows/automatically-categorize-and-label-gmail-emails-with-openai-gpt-4-classification-7633

Automatically Categorize and Label Gmail Emails with OpenAI GPT-4 Classification

1. Workflow Overview

This workflow automates the process of categorizing and labeling Gmail emails using OpenAI's GPT-5 model. It targets users or organizations aiming to intelligently organize incoming emails by priority and type, reducing manual inbox management effort.

The workflow consists of the following logical blocks:

1.1 Scheduled Trigger and Email Fetching: Periodically triggers the workflow and fetches recent emails from Gmail.
1.2 Email Processing Loop: Iterates over each fetched email individually for processing.
1.3 AI-Based Classification: Uses OpenAI GPT-5 to classify each email into predefined categories based on subject and body content.
1.4 Label Application: Applies Gmail labels corresponding to the AI classification result.
1.5 Workflow Control Loop: Ensures all emails are processed by looping back after labeling.

2. Block-by-Block Analysis

1.1 Scheduled Trigger and Email Fetching

Overview:
This block periodically triggers the workflow (every 5 minutes) and fetches up to 10 recent emails received within the last 5 minutes from Gmail.

Nodes Involved:

Schedule Trigger
Gmail - Get Emails
Sticky Note6

Node Details:

Schedule Trigger
- Type: Schedule Trigger node
- Configuration: Interval set to trigger every 5 minutes (field: minutes).
- Input: None (start node)
- Output: Triggers Gmail email fetching node.
- Potential Failures: Scheduling misconfiguration, node downtime.
Gmail - Get Emails
- Type: Gmail node (Get All operation)
- Configuration:
  - Limit: 10 emails
  - Filter: Received after current time minus 50000 minutes (likely a misconfiguration; intended to be 5 minutes but set as 50000)
  - Credentials: OAuth2 for Gmail account "Billy Email 2"
- Input: Triggered by Schedule Trigger
- Output: Passes fetched emails to Loop Over Items node
- Edge Cases:
  - API rate limits or authentication errors
  - Empty email fetch if no new emails
  - Potential misfiltering due to incorrect filter time (50000 minutes ≈ 34 days)
- Notes: The filter should be corrected to 5 minutes for intended functionality.
Sticky Note6
- Type: Sticky Note
- Content: Advises on configuring Gmail email fetching parameters such as time range and email limit.

1.2 Email Processing Loop

Overview:
Splits the batch of fetched emails into individual items to process each email separately.

Nodes Involved:

Loop Over Items

Node Details:

Loop Over Items
- Type: Split In Batches node
- Configuration: Default batch size (processes one email at a time)
- Input: Receives array of emails from Gmail - Get Emails
- Output: Sends single email record to Label Classifier node
- Edge Cases:
  - Large email batches could increase processing time
  - Failure if input data is empty or malformed

1.3 AI-Based Classification

Overview:
Classifies each email's subject and body into one of three categories (High Priority, Personal, Promotions) using OpenAI GPT-5 model via a LangChain text classifier.

Nodes Involved:

Label Classifier
OpenAI Chat Model
Sticky Note5

Node Details:

Label Classifier
- Type: LangChain Text Classifier node
- Configuration:
  - Input Text: Concatenation of email subject and body from JSON fields headers.subject and text
  - Categories:
    - High Priority: Emails requiring immediate attention
    - Personal: Non-work, private matters
    - Promotions: Marketing or advertisement messages
  - Fallback category: "other"
- Input: Receives one email from Loop Over Items
- Output: Routes email for labeling based on classification
- Edge Cases:
  - Misclassification due to ambiguous or malformed email text
  - Failure if email body or subject fields are missing
- Notes: Categories and descriptions are configurable to suit user needs.
OpenAI Chat Model
- Type: LangChain OpenAI Chat Model node
- Configuration: Uses GPT-5 model (latest OpenAI language model)
- Credentials: OpenAI API key
- Role: Supplies AI model for text classification node
- Input/Output: Connected only as AI language model provider to Label Classifier
- Edge Cases:
  - API rate limiting or authentication failures
  - Timeout or unexpected API errors
Sticky Note5
- Explains that the label classifier categories should be adjusted and described for best results.

1.4 Label Application

Overview:
Applies Gmail labels to each email thread according to the AI classification result. Labels include "IMPORTANT," "CATEGORY_PERSONAL," and "CATEGORY_PROMOTIONS."

Nodes Involved:

Add Label (High Priority)
Add Label (Personal)
Add Label (Promotions)
Sticky Note4

Node Details:

Add Label (High Priority), Add Label (Personal), Add Label (Promotions)
- Type: Gmail node (Add Labels operation on thread resource)
- Configuration:
  - Label IDs correspond to Gmail system or custom labels: "IMPORTANT," "CATEGORY_PERSONAL," "CATEGORY_PROMOTIONS"
  - Uses threadId from the email JSON to target the correct email thread
  - Credentials: Gmail OAuth2 for "Billy Email 2" account
- Input: Receives emails filtered by classification from Label Classifier
- Output: Returns to Loop Over Items node to process next email
- Edge Cases:
  - Failure to add label due to invalid label ID or Gmail API errors
  - Authentication expiration
  - Thread ID missing or invalid
- Notes: User should add or remove label nodes based on their label setup.
Sticky Note4
- Advises users to add or remove label nodes according to the number and names of their Gmail labels.

1.5 Workflow Control Loop

Overview:
After labeling an email, the workflow loops back to process the next email in the batch until all are handled.

Nodes Involved:

Loop Over Items (loop back connections)
Label Classifier (loop back connections)
Label application nodes (loop back connections)

Node Details:

Loop connections between Label Classifier, Add Label nodes, and Loop Over Items ensure continuous processing of email batches.
Edge Cases: Infinite loop risk if node outputs are misconfigured; ensure batch completes.

3. Summary Table

Node Name	Node Type	Functional Role	Input Node(s)	Output Node(s)	Sticky Note
Schedule Trigger	Schedule Trigger	Periodic workflow trigger	None	Gmail - Get Emails
Gmail - Get Emails	Gmail (Get All)	Fetch recent emails	Schedule Trigger	Loop Over Items	Advises on configuring email fetch parameters (time range, limit)
Loop Over Items	Split In Batches	Process emails one-by-one	Gmail - Get Emails	Label Classifier
Label Classifier	LangChain Text Classifier	Classify email content	Loop Over Items	Add Label (High Priority), Add Label (Personal), Add Label (Promotions), Loop Over Items	Advises to adjust categories and descriptions for best classification results
OpenAI Chat Model	LangChain OpenAI Chat Model	Provides GPT-5 model for classification	None (AI model provider)	Label Classifier
Add Label (High Priority)	Gmail (Add Labels on Thread)	Apply "IMPORTANT" label to email	Label Classifier	Loop Over Items	Advises to add/remove label nodes based on user labels
Add Label (Personal)	Gmail (Add Labels on Thread)	Apply "CATEGORY_PERSONAL" label	Label Classifier	Loop Over Items	Advises to add/remove label nodes based on user labels
Add Label (Promotions)	Gmail (Add Labels on Thread)	Apply "CATEGORY_PROMOTIONS" label	Label Classifier	Loop Over Items	Advises to add/remove label nodes based on user labels
Sticky Note	Sticky Note	Setup instructions	None	None	Workflow Configurations and required credentials
Sticky Note1	Sticky Note	Workflow process overview	None	None	Overview of workflow steps
Sticky Note2	Sticky Note	Workflow purpose and capabilities	None	None	Describes automation benefits and AI classification approach
Sticky Note4	Sticky Note	Label node customization advice	None	None	Advises on adding label nodes based on label count
Sticky Note5	Sticky Note	Label classifier category advice	None	None	Advises to adjust categories and descriptions
Sticky Note6	Sticky Note	Gmail get emails configuration notes	None	None	Advises on email fetch settings

4. Reproducing the Workflow from Scratch

Create Schedule Trigger Node
- Type: Schedule Trigger
- Configure to run every 5 minutes (interval: minutes, value: 5).
Create Gmail - Get Emails Node
- Type: Gmail node
- Operation: Get All
- Set limit to 10 emails
- Set filter: Received after current time minus 5 minutes
  - Use expression: ={{ $now.minus({ minutes: 5 }).toMillis() }}
- Connect Schedule Trigger output to this node.
- Attach Gmail OAuth2 credentials (setup with Gmail account).
Create Loop Over Items Node
- Type: Split In Batches
- Default batch size (process one email at a time)
- Connect Gmail - Get Emails output to this node input.
Create OpenAI Chat Model Node
- Type: LangChain OpenAI Chat Model
- Model: GPT-5
- Provide OpenAI API credentials.
Create Label Classifier Node
- Type: LangChain Text Classifier
- Input Text:
```
Email Subject:
{{ $json.headers.subject }}

Email Body:
{{ $json.text }}
```
- Define categories:
  - High Priority: Emails that require immediate attention or urgent action.
  - Personal: Personal or non-work related matters.
  - Promotions: Marketing or advertising content.
- Set fallback category to "other".
- Connect Loop Over Items output to this node input.
- Connect OpenAI Chat Model node as AI language model provider.
Create Add Label Nodes for Each Category
For each category, create a Gmail node configured to add the corresponding label to the email thread:
- Node Name: Add Label (High Priority)
  - Label ID: IMPORTANT
  - Resource: thread
  - Thread ID: ={{ $json.threadId }}
  - Operation: addLabels
  - Connect Label Classifier's "High Priority" output to this node.
  - Attach Gmail OAuth2 credentials.
- Node Name: Add Label (Personal)
  - Label ID: CATEGORY_PERSONAL
  - Same configuration as above.
  - Connect Label Classifier's "Personal" output.
- Node Name: Add Label (Promotions)
  - Label ID: CATEGORY_PROMOTIONS
  - Same configuration as above.
  - Connect Label Classifier's "Promotions" output.
Connect Each Add Label Node Back to Loop Over Items
- This ensures the workflow loops over remaining emails until all are processed.
Add Sticky Notes for Documentation (optional but recommended)
- Add setup instructions about configuring labels and credentials.
- Add workflow overview and purpose.
- Add notes about adjusting classifier categories and labels.
Credential Setup
- Gmail OAuth2: Configure with Gmail account scopes allowing reading emails and modifying labels.
- OpenAI API Key: Configure with access to GPT-5 or GPT-4 (depending on availability).
Test Workflow
- Run the workflow manually or wait for scheduled trigger.
- Monitor logs for errors or API limits.
- Adjust time filter and batch size as needed.

5. General Notes & Resources

Note Content	Context or Link
Workflow automatically classifies emails every 5 minutes using GPT-5 for intelligent inbox sorting.	Workflow description
Update label names and classifier categories as per your Gmail labels and business needs.	Sticky Note4 and Sticky Note5
Gmail OAuth2 credentials require proper scopes for reading emails and modifying labels.	Gmail OAuth2 credential configuration
OpenAI API key must have access rights to GPT-5 or GPT-4 models used for text classification.	OpenAI API credential configuration
Make sure to adjust the time filter in Gmail node to match your email fetch window (e.g., last 5 minutes).	Sticky Note6 and node configuration
This workflow is designed to be modular: you can add more label nodes for additional categories.	General workflow design
For more info on n8n Gmail node: https://docs.n8n.io/integrations/builtin/app-nodes/n8n-nodes-base.gmail/	Official n8n documentation
For more info on LangChain text classifiers in n8n: https://docs.n8n.io/integrations/builtin/app-nodes/langchain/	Official n8n documentation

Disclaimer: The provided text derives exclusively from an automated workflow created with n8n, a tool for integration and automation. This process strictly complies with applicable content policies and contains no illegal, offensive, or protected material. All data processed is legal and public.

16 KiB Raw Blame History

Automatically Categorize and Label Gmail Emails with OpenAI GPT-4 Classification

1. Workflow Overview

2. Block-by-Block Analysis

1.1 Scheduled Trigger and Email Fetching

1.2 Email Processing Loop

1.3 AI-Based Classification

1.4 Label Application

1.5 Workflow Control Loop

3. Summary Table

4. Reproducing the Workflow from Scratch

5. General Notes & Resources

16 KiB

Raw Blame History