lab/nusquama-n8nworkflows-xyz

Fork 0

mirror of https://github.com/nusquama/n8nworkflows.xyz.git synced 2025-11-27 01:35:37 +08:00

Files

nusquama 5d7083d7d7 creation

2025-11-14 10:39:58 +01:00

19 KiB

Raw Blame History

Process Resumes from Google Drive to ClickUp using GPT-4o and Google Sheets

https://n8nworkflows.xyz/workflows/process-resumes-from-google-drive-to-clickup-using-gpt-4o-and-google-sheets-8529

Process Resumes from Google Drive to ClickUp using GPT-4o and Google Sheets

1. Workflow Overview

This workflow automates the processing of newly uploaded resumes in a designated Google Drive folder, extracting structured candidate data using GPT-4o AI, storing that data in Google Sheets, and creating corresponding hiring tasks in ClickUp. It is designed to streamline recruitment pipelines by eliminating manual data entry and ensuring rapid, consistent candidate information handling.

The workflow is logically divided into these blocks:

1.1 Input Reception: Detect new resume files uploaded to Google Drive.
1.2 File Processing: Download the resume file and convert from PDF to raw text.
1.3 AI Data Extraction: Use GPT-4o to analyze the resume text and extract key candidate details as JSON.
1.4 Data Cleaning: Sanitize and verify the AI output JSON for database compatibility.
1.5 Database Storage: Append or update the candidate information in a Google Sheets database.
1.6 Task Creation: Generate a new hiring task in ClickUp assigned to the recruitment team.

2. Block-by-Block Analysis

1.1 Input Reception

Overview:
This block continuously monitors a specific Google Drive folder for new resume files. Upon detecting a new file creation event, it triggers the workflow to start processing.

Nodes Involved:

Watch for New Resumes
Sticky Note - Trigger (documentation node)

Node Details:

Watch for New Resumes
- Type: Google Drive Trigger
- Role: Monitors a folder on Google Drive for new files (event = fileCreated)
- Configuration: Watches folder with ID 1MIvpHU_ZqG76Vov2-D5WlS5dD3UhOMSz named "Resume_store", polls every minute
- Inputs: None (trigger node)
- Outputs: Emits metadata about new files (including file ID)
- Edge cases: Potential delays due to polling frequency, Google Drive API rate limits, auth token expiration
- Sticky Note Reference: "🚀 WORKFLOW START" explains trigger settings and user tips

1.2 File Processing

Overview:
Downloads the detected resume file from Google Drive and converts it from PDF format into plain text suitable for AI processing.

Nodes Involved:

Download Resume File
Convert PDF to Text
Sticky Note - Download
Sticky Note - Extract

Node Details:

Download Resume File
- Type: Google Drive
- Role: Downloads the actual resume file (PDF) given the file ID from trigger
- Configuration: Operation = download; File ID dynamically taken from trigger output
- Inputs: File metadata from "Watch for New Resumes"
- Outputs: Binary PDF file data
- Edge cases: File not found, permission errors, network timeouts
Convert PDF to Text
- Type: Extract From File
- Role: Extracts text content from PDF binary data
- Configuration: Operation = pdf extraction
- Inputs: Binary PDF data from download node
- Outputs: Text content of resume
- Edge cases: Scanned image PDFs will have poor extraction results; corrupted PDFs may fail
Sticky Notes
- Provide detailed explanations of these steps and tips for users
- Highlight reliance on text-based PDFs for best extraction quality

1.3 AI Data Extraction

Overview:
Uses an advanced GPT-4o AI model via Langchain agent to parse the resume text and extract structured candidate fields in JSON format.

Nodes Involved:

AI Resume Analyzer (Langchain Agent)
AI Language Model (Azure OpenAI GPT-4o-mini)
Sticky Note - AI Analysis

Node Details:

AI Resume Analyzer
- Type: Langchain Agent
- Role: Defines the prompt and calls the AI language model to process resume text
- Configuration: Prompt instructs AI to extract fields — name, email, phone, experience years, skills array, current role, education — from resume text (passed dynamically)
- Inputs: Resume text string from PDF conversion
- Outputs: Raw AI JSON response string (wrapped in code fences)
- Edge cases: AI may produce malformed JSON, partial data, or hallucinated info; requires downstream cleaning
AI Language Model
- Type: Langchain LM Chat Azure OpenAI
- Role: Runs the GPT-4o-mini model to generate AI response
- Configuration: Model set to gpt-4o-mini, no additional options
- Credentials: Azure OpenAI API account
- Inputs: Prompt from AI Resume Analyzer
- Outputs: AI-generated text response
- Edge cases: API rate limits, authentication failures, timeouts
Sticky Note
- Details AI role and fields extracted
- Emphasizes AI’s ability to handle diverse resume formats

1.4 Data Cleaning

Overview:
Processes the AI raw JSON output to remove formatting artifacts and parse it into a clean JSON object suitable for database insertion.

Nodes Involved:

Clean AI Response (Code node)
Sticky Note - Clean Data

Node Details:

Clean AI Response
- Type: Code (JavaScript)
- Role: Strips out markdown code fences (json), trims text, attempts safe JSON parse
- Configuration: Custom JS code that catches JSON parse errors and returns error info if parsing fails
- Inputs: Raw AI JSON string from AI Resume Analyzer
- Outputs: Parsed JSON object or error object
- Edge cases: Malformed AI output causing parse failure; returns error message for manual review
- Sub-workflow: None
Sticky Note
- Explains cleaning steps and rationale
- Recommends validation before saving data

1.5 Database Storage

Overview:
Appends or updates the cleaned candidate details in a centralized Google Sheets document acting as the candidate database.

Nodes Involved:

Save Candidate to Database (Google Sheets)
Sticky Note - Database

Node Details:

Save Candidate to Database
- Type: Google Sheets
- Role: Append or update candidate record based on matching candidate name
- Configuration:
  - Document ID: Google Sheet with ID 1JlXxy90s0we_IqErHyvomrJSijb8pd4H91hOUCH6xCA
  - Sheet: "Sheet1" (gid=0)
  - Operation: appendOrUpdate (match on "Name" column with $json.name)
  - Fields saved: Email, Phone, Years of experience, Skills, Current role, Education
- Inputs: Cleaned JSON with candidate fields
- Outputs: Confirmation of sheet update
- Credentials: Google Sheets OAuth2
- Edge cases: Duplicate names causing overwrite; Google API quota limits; auth expiration
Sticky Note
- Describes database role and data stored
- Highlights permanent record keeping and searchability

1.6 Task Creation

Overview:
Creates a new task in ClickUp corresponding to the candidate, assigning it to the hiring team to initiate the recruitment process.

Nodes Involved:

Create Hiring Task (ClickUp)
Sticky Note - Task Creation

Node Details:

Create Hiring Task
- Type: ClickUp
- Role: Creates a new task titled “[Candidate Name] Hiring process” in a specific ClickUp list
- Configuration:
  - Team: 9016683627
  - Space: 90162844741
  - Folder: 90164394824
  - List: 901610812551
  - Task name: dynamic from $json.name + " Hiring process"
  - Assignees: user ID 95074494 (recruiting team member)
- Inputs: Candidate name from Google Sheets output
- Outputs: Created task metadata
- Credentials: ClickUp API OAuth2
- Edge cases: API failure, invalid team/list IDs, permission errors
Sticky Note
- Details task creation, assignee, and its role in recruitment pipeline

3. Summary Table

Node Name	Node Type	Functional Role	Input Node(s)	Output Node(s)	Sticky Note
Watch for New Resumes	Google Drive Trigger	Detect new resume uploads	None	Download Resume File	🚀 WORKFLOW START: Monitors folder Resume_store every 1 minute for new files
Download Resume File	Google Drive	Download resume PDF file	Watch for New Resumes	Convert PDF to Text	📥 STEP 1: Downloads resume PDF from Drive
Convert PDF to Text	Extract From File	Extract text from PDF	Download Resume File	AI Resume Analyzer	📄 STEP 2: Converts PDF to text, best with text-based PDFs
AI Resume Analyzer	Langchain Agent	Extract candidate info using AI	Convert PDF to Text	Clean AI Response	🤖 STEP 3: Extracts structured candidate fields from resume text using GPT-4o
AI Language Model	Langchain LM Chat Azure OpenAI	Runs GPT-4o-mini model	AI Resume Analyzer	AI Resume Analyzer	(No sticky note)
Clean AI Response	Code (JavaScript)	Clean and parse AI JSON output	AI Resume Analyzer	Save Candidate to Database	🔧 STEP 4: Removes code fences, safely parses JSON, handles errors
Save Candidate to Database	Google Sheets	Store/update candidate in sheet	Clean AI Response	Create Hiring Task	💾 STEP 5: Saves candidate data to Google Sheets, updates existing records
Create Hiring Task	ClickUp	Create hiring task for candidate	Save Candidate to Database	None	✅ STEP 6 - FINAL: Creates new ClickUp task assigned to hiring team
Sticky Note - Trigger	Sticky Note	Documentation	None	None	See details in node
Sticky Note - Download	Sticky Note	Documentation	None	None	See details in node
Sticky Note - Extract	Sticky Note	Documentation	None	None	See details in node
Sticky Note - AI Analysis	Sticky Note	Documentation	None	None	See details in node
Sticky Note - Clean Data	Sticky Note	Documentation	None	None	See details in node
Sticky Note - Database	Sticky Note	Documentation	None	None	See details in node
Sticky Note - Task Creation	Sticky Note	Documentation	None	None	See details in node
Sticky Note - Overview	Sticky Note	Documentation	None	None	Overview of entire workflow purpose and benefits

4. Reproducing the Workflow from Scratch

Create a Google Drive Trigger node:
- Type: Google Drive Trigger
- Configure to watch folder ID 1MIvpHU_ZqG76Vov2-D5WlS5dD3UhOMSz ("Resume_store")
- Event: fileCreated
- Polling interval: every 1 minute
- Credential: Google Drive OAuth2 account
Create a Google Drive node to download file:
- Type: Google Drive
- Operation: download
- File ID: Reference from trigger output ({{$json.id}})
- Credential: Google Drive OAuth2 account
- Connect output of trigger node here
Create an Extract From File node:
- Type: Extract From File
- Operation: pdf
- Input: Binary data from download node
- Connect output of download node here

Create a Langchain Agent node (AI Resume Analyzer):

Type: Langchain Agent
Prompt Type: define

Text prompt to AI:

Extract the following fields from the resume text below and return them as JSON only:

{
  "name": "",
  "email": "",
  "phone": "",
  "experience_years": "",
  "skills": [],
  "current_role": "",
  "education": ""
}

Resume text:
{{$json["text"]}}

Connect output of PDF text extraction node here

Create a Langchain LM Chat Azure OpenAI node:
- Type: Langchain LM Chat Azure OpenAI
- Model: gpt-4o-mini
- Credential: Azure OpenAI API account
- Connect output of Langchain Agent node here
Connect Langchain LM node output back to Langchain Agent node’s AI Language Model input:
- This connection enables Langchain Agent to use the LM node for processing

Create a Code node (Clean AI Response):

Type: Code (JavaScript)

Code:

return items.map(item => {
  let text = item.json.output;
  text = text.replace(/```json|```/g, "").trim();
  let parsed = {};
  try {
    parsed = JSON.parse(text);
  } catch (err) {
    parsed = { error: "Failed to parse JSON", raw: text };
  }
  return { json: parsed };
});

Connect output of AI Resume Analyzer here

Create a Google Sheets node (Save Candidate to Database):
- Type: Google Sheets
- Operation: appendOrUpdate
- Document ID: 1JlXxy90s0we_IqErHyvomrJSijb8pd4H91hOUCH6xCA
- Sheet Name: "Sheet1"
- Value To Match On: {{$json.name}}
- Column To Match On: Name
- Fields to save (mapped from input JSON):
  - Email → Email
  - Phone → Phone
  - Years of experience → experience_years
  - Skills → skills
  - Current role → current_role
  - Education → education
- Credential: Google Sheets OAuth2 account
- Connect output of Clean AI Response node here
Create a ClickUp node (Create Hiring Task):
- Type: ClickUp
- Team ID: 9016683627
- Space ID: 90162844741
- Folder ID: 90164394824
- List ID: 901610812551
- Task Name: {{$json.name}} Hiring process
- Additional Fields: Assignees → [95074494]
- Credential: ClickUp API OAuth2 account
- Connect output of Google Sheets node here
Set up sticky notes for documentation at each key step (optional but recommended):
- Describe purpose and user tips for each block (trigger, download, extract, AI, clean, save, task creation).
- Use markdown for clarity.
Validate credentials and API permissions:
- Google Drive: Folder access and file read
- Google Sheets: Edit permissions on the candidate database sheet
- Azure OpenAI: Access to GPT-4o-mini model
- ClickUp: Permissions to create tasks in specified workspace and assign members
Test workflow:
- Upload a sample PDF resume to the watched Google Drive folder
- Observe automatic processing, data extraction, database update, and task creation

5. General Notes & Resources

Note Content	Context or Link
To ensure accurate text extraction, resumes should be text-based PDFs rather than scanned images. OCR is not implemented in this workflow.	Sticky Note - Extract
The AI model prompt is crafted to return JSON only, but occasional malformed outputs are handled gracefully by the cleaning code node.	Sticky Note - AI Analysis, Clean AI Response
The Google Sheets database acts as a central, searchable HR repository updated automatically to prevent duplicate entries based on candidate name matching.	Sticky Note - Database
ClickUp task names follow the convention “[Candidate Name] Hiring process” and are automatically assigned to a recruiting team member with user ID 95074494.	Sticky Note - Task Creation
Workflow runtime is under 2 minutes per resume, significantly accelerating recruitment pipelines by automating repetitive tasks.	Sticky Note - Overview
Requires setup of OAuth2 credentials for Google Drive, Google Sheets, ClickUp, and Azure OpenAI in n8n prior to deployment.	Workflow prerequisites
For troubleshooting, monitor API quotas and error logs, especially for AI parsing errors and Google API rate limits.	Best practice note
More info on ClickUp API: https://clickup.com/api; Google Sheets API: https://developers.google.com/sheets/api; Azure OpenAI: https://learn.microsoft.com/en-us/azure/cognitive-services/openai/	Useful external resource links

This structured document fully explains the workflow’s design, nodes, configuration, and operational logic, enabling efficient reproduction, debugging, and extension by advanced users or automated agents.

19 KiB Raw Blame History Unescape Escape

Process Resumes from Google Drive to ClickUp using GPT-4o and Google Sheets

1. Workflow Overview

2. Block-by-Block Analysis

1.1 Input Reception

1.2 File Processing

1.3 AI Data Extraction

1.4 Data Cleaning

1.5 Database Storage

1.6 Task Creation

3. Summary Table

4. Reproducing the Workflow from Scratch

5. General Notes & Resources

19 KiB

Raw Blame History