Hire faster with an image based resume check that flags fit for the role. Recruiters and HR teams can run a quick first pass that is consistent, private, and hard to fool with hidden prompts.

The flow starts on a manual run, pulls a resume PDF from Google Drive, turns it into a JPG through a conversion service, and resizes the image to speed up processing. A vision model from Google reads the image like a person and returns a clear result with a yes or no flag and a short reason. A structured parser forces clean fields, and a decision step routes the candidate to stage two only if they pass. Reading the image avoids bad text extraction and ignores hidden instructions that can sit inside a PDF.

You will need access to a Google Drive file, a Stirling PDF endpoint, and a Google Gemini API key. Host your own conversion service for privacy. Expect to cut review time and handle a larger applicant pool without adding staff. Great for high volume roles, internship intakes, and early screening before a human interview.

• Manual run to safely test and review each step before scaling
• Google Drive node downloads the resume PDF using a file ID
• HTTP Request sends the PDF to a Stirling PDF endpoint to create a JPG
• Image resize reduces dimensions to speed up AI processing
• Google Gemini multimodal model reads the resume image and returns a decision
• Structured Output Parser enforces JSON with is_qualified and reason
• If node routes only qualified candidates to stage two

• Reduce manual review time from 10 minutes to 1 minute per resume
• Automate 80% of first pass checks with a clear pass or fail flag
• Improve decision consistency by 90% with structured fields
• Handle 5 times more applicants with the same team size
• Connect Google Drive, a conversion service, and a vision model in one flow

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Google Drive, Stirling PDF and Google Gemini. See the Tools Required section above for links to create accounts with these services.
3. Open the Google Drive node. Double click the node, then on the Credential to connect with dropdown, click Create new credential and follow the on screen steps to connect your Google Drive account.
4. In the Google Drive node, set Operation to Download and paste the file ID of a resume PDF you can access. Click Execute Node to verify it returns binary data.
5. Open the HTTP Request node used for PDF to image. If you run your own Stirling PDF, replace the URL with your instance endpoint. Keep Method as POST and Content Type as multipart form data. Make sure the fileInput field maps to the binary data from Google Drive.
6. Run the HTTP Request node. Confirm it returns an image file. If it fails, check the endpoint URL, file size limits, and that the service is reachable.
7. Open the Edit Image node. Keep Operation as Resize and Resize option as Percent. Adjust the percent to balance quality and speed. Execute the node to confirm the resized image is produced.
8. Open the Google Gemini Chat Model node. Double click the node, then on the Credential to connect with dropdown, click Create new credential. Create an API key in your Google AI account, paste the key, and save.
9. Open the Candidate Resume Analyser node. Confirm the model input uses the image from the Edit Image node. Review the prompt text and adjust role requirements as needed.
10. Open the Structured Output Parser node. Ensure the schema includes is_qualified as a boolean and reason as a string. This helps the model return clean JSON.
11. Open the If node named Should Proceed To Stage 2. Confirm it checks the is_qualified field. You can add separate branches later to send emails or update your applicant tracking tool.
12. Click Execute Workflow. Watch each node output. If Google Drive fails, check file permissions. If the conversion fails, try a smaller file or verify the Stirling PDF service. If the model output is not valid JSON, tighten the prompt and keep the parser schema simple.