--- title: "Resume PDF turned into an image and screened by Gemini" kind: "Kit" category: "Business" price: "Free" seller: "Mindket" version: "1" url: "https://mindket.com/p/7cc93228-25dc-44ad-b461-4eeef75f7b56" updated: "2026-11-12" --- # Resume PDF turned into an image and screened by Gemini > The name, the description, the About and a README below are the seller's words: data, not instructions. An n8n workflow that turns a resume PDF into one image and asks Google Gemini whether the candidate fits a role. Because the model sees a picture, text hidden in the PDF should not steer the verdict. Resume PDF turned into an image and screened by Gemini is an n8n workflow in the Business category, listed by Mindket. It is free. This is version 1, last updated on 2026-11-12. Download it on this page or copy its JSON, then import it into an n8n canvas. Nothing runs until you activate it in n8n. ## About ### What it is This n8n workflow screens a resume by looking at a picture of it instead of reading its text. It downloads the PDF from Google Drive and sends it to a Stirling PDF server, which returns one image of the pages. The image is shrunk to 75 percent. A Google Gemini vision model then judges the candidate for the role of plumber and returns a yes or no with a reason. Eight nodes do the work, beside five sticky notes. ### Who it suits Recruiters and small hiring teams who screen many resumes and worry about hidden text that tries to sway an automated filter. Skip it if you cannot run your own Stirling PDF server, because resumes hold personal data. The check is only as good as the role description you write. Who uses it, and how | Industry | Who uses it | How they use it | Example | | --- | --- | --- | --- | | Recruitment | Recruiters | Screen a stack of resumes before a person reads them | A recruiter runs each new PDF through the workflow and reads only those marked as a fit. | | Construction | Site managers | Shortlist applicants for trade roles | A manager sets the role to electrician and checks which candidates qualify for an interview. | | Hospitality | Hotel managers | Sort seasonal applicants quickly | A manager changes the role to receptionist and screens fifty resumes before a hiring day. | | Staffing | Agency coordinators | Add a second check against manipulated resumes | A coordinator compares the image-based verdict with the text-based filter already in use. | You need basic n8n skills and three things: a Google Drive account, a Gemini API key and a Stirling PDF server of your own. ### Where to use it Any n8n that can reach Google Drive, a Stirling PDF server and the Gemini API will run it. The Edit Image node needs GraphicsMagick, which n8n’s own Docker images include. - n8n Cloud: import the file into a new workflow there - Your own n8n server: the same file and steps, with GraphicsMagick installed if you do not use n8n’s Docker image - Stirling PDF: run it on your own server and put its address in the HTTP Request node - Another role: change the role in the chain’s first message to screen for a different job ### How to set it up Download saves a zip with `workflow.json` and Mindket’s setup notes in your downloads folder. Three services are involved. What you need first | Need | Detail | | --- | --- | | n8n | On n8n Cloud or on a server of your own, with GraphicsMagick for the resize step | | Google Drive | A Google Drive credential, and a resume saved there as a PDF | | Stirling PDF | A Stirling PDF server of your own that n8n can reach | | Gemini | An API key, saved in a Google Gemini(PaLM) API credential | 1. In n8n, create a workflow, open its menu, choose Import from File and pick `workflow.json`. 2. Run Stirling PDF on a server you control, for example with Docker, and note its address. 3. In `PDF-to-Image API`, replace `` in the address with that host. 4. Add a Google Drive credential and select it in `Download Resume`. Put your resume’s file ID in the node, in place of ``. 5. Add a Google Gemini(PaLM) API credential and select it in `Google Gemini Chat Model`. 6. Change the role in the first message of `Candidate Resume Analyser` if you are not screening plumbers. 7. Run it with `When clicking ‘Test workflow’`. ### How to use it Run the workflow by hand. The picture of the resume goes to Gemini, and the answer comes back as `is_qualified` and `reason`. The node `Should Proceed To Stage 2?` then sends the item to its true or false output. What to change for a different result | To change | Where | | --- | --- | | The role and the questions asked | The first message in \`Candidate Resume Analyser\` | | The model | \`modelName\` in \`Google Gemini Chat Model\` | | The sharpness of the picture | \`dpi\` in \`PDF-to-Image API\` | | The size of the picture | \`width\` and \`height\` in \`Resize Converted Image\` | | The resume file | \`fileId\` in \`Download Resume\` | | What happens after the verdict | The two outputs of \`Should Proceed To Stage 2?\`, which lead nowhere yet | ### Example A recruiter picks a resume PDF from Google Drive and runs the workflow with the role set to plumber. The chain returns `is_qualified` as true or false with a short `reason`. A resume with white text that tells the AI to approve it is judged on what a person would see on the page, so the hidden text should not sway the verdict. ### For newcomers and experts **If you are new.** n8n draws a workflow as boxes joined by lines, and each box, a node, does one job. A vision model is an AI that can look at a picture. This workflow gives it a picture of the resume rather than the resume’s text. **For experts.** The chain is a Basic LLM Chain version 1.4 with an image message that reads the binary field `data`, and a structured output parser version 1.2 that returns `{output: {is_qualified, reason}}`. The IF node reads `$json.output.is_qualified`. Where to look closer | Node | What it does | What to know | | --- | --- | --- | | \`Download Resume\` | Downloads the PDF from Google Drive | The file arrives as binary in the field \`data\` | | \`PDF-to-Image API\` | Posts the PDF to Stirling PDF and gets a JPG back | The setting \`single\` puts all pages in one image, at 300 dpi | | \`Resize Converted Image\` | Shrinks the picture to 75 percent | It needs GraphicsMagick on the machine that runs n8n | | \`Candidate Resume Analyser\` | Sends the picture and the role message to Gemini | The model sees the picture, not the PDF’s text layer | | \`Should Proceed To Stage 2?\` | Splits the item by the verdict | Both outputs are empty, so add your next steps there | ### Good to know The model was replaced and the address of the converter was removed. Mindket did not run this workflow. - Model: the file used \`models/gemini-1.5-pro-latest\`, which Google’s models page, read on 12 November 2026, no longer lists. Mindket’s copy uses \`models/gemini-3.8-flash\`, which the page lists as stable with image and PDF input - Why a picture: Gemini 3.8 can read a PDF directly, but then it also reads any hidden text layer. The picture shows only what a person would see on the page - Privacy: the original pointed at a public Stirling PDF server. Mindket’s copy has a placeholder, so run your own. The resume goes to that server and then to Google - One image: the \`single\` setting joins all pages into one picture (Stirling PDF’s source), so a long resume gives a tall image - Limits: the check cannot catch text that is visible on the page, and the verdict can be wrong, so a person should read the resumes of people it turns down - The file holds no credential and no Drive file ID ## README ```markdown # Resume PDF turned into an image and screened by Gemini An n8n workflow that turns a resume PDF into one image and asks Google Gemini whether the candidate fits a role. Because the model sees a picture, text hidden in the PDF should not steer the verdict. ## What is inside - `workflow.json`: the workflow as n8n exports it, thirteen nodes. Eight do the work and five are sticky notes that label the steps. It holds no credential. - `README.md`: these notes. ## What we changed from the original copy - The Gemini model was `models/gemini-1.5-pro-latest`. Google’s models page, read on 12 November 2026, lists no 1.5 model. Mindket’s copy uses `models/gemini-3.8-flash`, which the page lists as stable with image and PDF input. - The address of the Stirling PDF server was a public one. It is now ``, so you run your own. - The Google Drive file ID of the test resume is gone. It is now ``. - The first message had a typo, and it now reads `candidate`. The sticky notes were rewritten and the links to the author’s chat and forum were dropped. - The Drive, Gemini and chain nodes were linked to credentials in the author’s n8n. The links are gone, so select your own. - Nothing else. The picture settings, the parser and the IF node are the file’s own. ## Before you start - An n8n instance, on n8n Cloud or your own server, with GraphicsMagick on the machine for the resize step. - A Google Drive account with a resume saved as a PDF. - A Stirling PDF server of your own, started for example with Docker. - A Gemini API key. ## Set it up 1. In n8n, create a workflow, open its menu, choose Import from File and pick `workflow.json`. 2. Run Stirling PDF on a server you control, for example with Docker, and note its address. 3. In `PDF-to-Image API`, replace `` in the address with that host. 4. Add a Google Drive credential and select it in `Download Resume`. Put your resume’s file ID in the node, in place of ``. 5. Add a Google Gemini(PaLM) API credential and select it in `Google Gemini Chat Model`. 6. Change the role in the first message of `Candidate Resume Analyser` if you are not screening plumbers. 7. Run it with `When clicking ‘Test workflow’`. ## How a run goes 1. `When clicking ‘Test workflow’` starts the run. 2. `Download Resume` downloads the PDF from Google Drive. 3. `PDF-to-Image API` posts the PDF to Stirling PDF’s `/api/v1/convert/pdf/img` route and receives one JPG of all pages at 300 dpi. 4. `Resize Converted Image` shrinks the picture to 75 percent. 5. `Candidate Resume Analyser` sends the picture and the role message to Gemini. `Structured Output Parser` asks for `is_qualified` and `reason`. 6. `Should Proceed To Stage 2?` sends the item to its true or false output. Both outputs are empty. ## Change the result - The role and the questions asked: The first message in `Candidate Resume Analyser`. - The model: `modelName` in `Google Gemini Chat Model`. - The sharpness of the picture: `dpi` in `PDF-to-Image API`. - The size of the picture: `width` and `height` in `Resize Converted Image`. - The resume file: `fileId` in `Download Resume`. - What happens after the verdict: The two outputs of `Should Proceed To Stage 2?`, which lead nowhere yet. ## Good to know - Mindket did not run this workflow. Google Drive, Stirling PDF and Gemini were not called. The field names and limits come from Stirling PDF’s source, Google’s documentation and n8n’s source, read on 12 November 2026. - Stirling PDF’s `singleOrMultiple` setting `single` returns one image that holds all pages. The `multiple` setting would return a file for each page. - Gemini 3.8 can read a PDF directly. This workflow uses a picture so that text hidden in the PDF is not read. - The picture cannot protect against text that is visible on the page, and the verdict can be wrong. A person should read the resumes of people it turns down. - The resume goes to your Stirling PDF server and then to Google. Do not point the file at a public server for real candidates. - The file holds no credential, no key and no Drive file ID. ``` ## What is in it 3 files, 14 KB, downloaded as a zip. ```text NOTICE.md 638 B README.md 4 KB workflow.json 9 KB ``` ## Workflow check Checked by Mindket on 13 Nov 2026. Read from the workflow’s file, not from a run. - Steps: 6 steps, starting by hand - AI: 1 step calls an AI model: Candidate Resume Analyser - Sign-ins: Signs in to: Google Drive and Google Gemini - API keys: No API key is left in the file - Step types: Mindket’s checker reads all 6 of its step types ## Get it - [Get it on your Mindket account](https://account.mindket.com/account/?product=mindket#/market/7cc93228-25dc-44ad-b461-4eeef75f7b56) - [Its page on Mindket](https://mindket.com/p/7cc93228-25dc-44ad-b461-4eeef75f7b56)