You finish a job, snap a photo, and then nothing happens. The photo sits in your camera roll while your business Facebook page collects dust — not because you don't have great work to show, but because writing captions takes time you don't have when the next job is already waiting. This weekend, you can fix that with a simple automation that does the first draft for you.
The Workflow in Plain English
The idea is straightforward: you drop finished-job photos into a dedicated Google Drive folder. An automation picks them up, sends each image to Google Gemini Flash for analysis, and writes a ready-to-review caption into a Google Sheet. You spend a few seconds approving and adjusting, instead of staring at a blank text box.
Here's the full flow:
- Trigger: a new photo lands in your Post Queue folder in Google Drive.
- Analyze: n8n sends the image to Gemini Flash, which looks at the photo and writes a short social media caption.
- Log: the caption and a link to the original photo get added as a new row in Google Sheets.
- Publish: you review the row, tweak anything that needs it, and post — manually or with a direct platform connection.
The key insight is that step 3 is your human checkpoint. The AI gives you something to react to instead of something to invent. That alone cuts most of the friction.
What You Need
Three things, all free to start:
- A Google account — you likely already have this. You will use Google Drive and Google Sheets, plus a free Gemini API key from Google AI Studio. Getting the key takes about two minutes.
- n8n — an automation tool with a built-in Google Gemini integration. You can use their cloud plan or run n8n yourself for free using Docker on a spare machine or a low-cost server. Their Google Gemini integration page lists exactly what the node supports.
- Your camera — get into the habit of uploading finished-job photos to the Drive folder while you are still on site, or at the end of each day.
If you prefer a no-code alternative to n8n, Make.com has a Google Drive and Gemini AI integration that works in the same pattern — trigger on new Drive file, call Gemini, write to Sheets.
Building It: Step by Step
Step 1: Create the folder. In Google Drive, make a folder called Post Queue. This is where every finished-job photo will land. If you use a phone, set up automatic upload or a shared folder your crew can drop photos into from the field.
Step 2: Get your Gemini API key. Go to Google AI Studio, sign in with your Google account, and generate an API key. Save it somewhere secure — you will paste it into n8n in a moment.
Step 3: Set up the n8n workflow. Start with a Google Drive trigger set to fire when a new file appears in your Post Queue folder. Connect it to a Google Drive — Download File node to pull the image data into the workflow.
The n8n template library has a ready-made starting point: Post Facebook image captions from Google Drive using Gemini Vision. Import it and adapt it rather than building from scratch.
Step 4: Write the caption prompt. In the Gemini node, pass the image and a prompt like this:
You are the social media manager for a local service business. Look at this photo of completed work and write a short, friendly social media caption (2-4 sentences). Mention the type of work if you can identify it. Tone: professional but approachable. End with a brief call to action like: Call us for a free estimate.
Adjust the prompt to match your voice. If you want three variations to choose from, ask for them explicitly: Write three versions — one formal, one conversational, one brief.
Step 5: Write to Google Sheets. Add a Google Sheets — Append Row node. Map the Gemini output to a Caption Draft column, and include a link to the original Drive file so you can see the photo alongside the caption. Add a Status column (Pending / Approved / Posted) for your own tracking.
Test the workflow with five or six real photos before calling it done. Adjust the prompt based on what comes back.
Practical Tips That Save You Headaches
- Start without auto-posting. The Sheets review step is not just a safety net — it is the step where you turn a generic AI description into something that sounds like your business. Most service businesses keep the review step permanently. It is worth it.
- Name your photos by job type. A filename like
fencing-cedar-2026-07.jpggives Gemini more context thanIMG_4812.jpg. You can pass the filename into the prompt so the AI is not guessing what it is looking at. - Include a brief brand note in your prompt. Add a sentence or two: Our business serves the Comox Valley and focuses on residential electrical work. Gemini uses it to write captions that are geographically grounded and match your scope of work.
- Add a quick notification step. After the Sheets row is created, add a simple email or Telegram message: New caption ready for review. A nudge on your phone beats remembering to check a spreadsheet.
- Wire up direct posting later. Facebook and Instagram both have Graph API integrations in n8n, but setting them up requires a Facebook Developer App and some OAuth configuration. Get the Sheets workflow running reliably first, then add direct publishing as a second weekend project once you trust the output.
What to Expect
Gemini Flash is good at identifying common work types — decks, rooflines, pipe runs, landscaping — and generating readable descriptions. It is weaker at capturing brand-specific details it cannot see in the photo, so context like your service area or specialty should go into the prompt itself, not just the image.
The goal is not a perfect caption written by a machine. It is a solid first draft waiting for you before you have even driven away from the job site. Most drafts will need one quick edit. Some will be ready to post as-is. Either way, you have replaced a blank-page problem with a thirty-second tweak.
Start with the Google Sheets output, spend a week posting from it, and then decide if you want to wire up direct publishing. The discipline of reviewing before posting is usually the right call — and it is the part that keeps your voice intact.