Google Sheets + LinkedIn, ready to review post drafts
Your LinkedIn “content plan” starts clean, then turns into a messy mix of half-read articles, copy-pasted notes, and drafts scattered across docs. And when you finally sit down to write, you’re basically starting from zero again.
This is where Sheets LinkedIn drafts automation helps. Social media managers feel it first, but founders and ghostwriters deal with the same bottleneck: too many topics, not enough time, and a feed that rewards consistency.
This workflow turns one “To do” topic in Google Sheets into a ready-to-review LinkedIn post draft plus an image URL, so you can approve, tweak, and publish with confidence. You’ll see how it works, what you need, and how to customize it without turning this into a technical project.
How This Automation Works
Here’s the complete workflow you’ll be setting up:
n8n Workflow Template: Google Sheets + LinkedIn, ready to review post drafts
flowchart LR
subgraph sg0["Start Workflow Flow"]
direction LR
n0@{ icon: "mdi:play-circle", form: "rounded", label: "Start Workflow", pos: "b", h: 48 }
n1@{ icon: "mdi:database", form: "rounded", label: " Get 'To do' Topics from Sheet", pos: "b", h: 48 }
n2["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/httprequest.dark.svg' width='40' height='40' /></div><br/>Scrape Articles via Dumpling.."]
n3@{ icon: "mdi:brain", form: "rounded", label: "Enable GPT-4o for LangChain ..", pos: "b", h: 48 }
n4@{ icon: "mdi:robot", form: "rounded", label: "Summarize 3 Articles + Gener..", pos: "b", h: 48 }
n5@{ icon: "mdi:swap-vertical", form: "rounded", label: " Extract Post Text and Image..", pos: "b", h: 48 }
n6["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/httprequest.dark.svg' width='40' height='40' /></div><br/>Generate Image with Dumpling.."]
n7@{ icon: "mdi:database", form: "rounded", label: "Update Google Sheet with Pos..", pos: "b", h: 48 }
n0 --> n1
n1 --> n2
n6 --> n7
n3 -.-> n4
n5 --> n6
n2 --> n4
n4 --> n5
end
%% Styling
classDef trigger fill:#e8f5e9,stroke:#388e3c,stroke-width:2px
classDef ai fill:#e3f2fd,stroke:#1976d2,stroke-width:2px
classDef aiModel fill:#e8eaf6,stroke:#3f51b5,stroke-width:2px
classDef decision fill:#fff8e1,stroke:#f9a825,stroke-width:2px
classDef database fill:#fce4ec,stroke:#c2185b,stroke-width:2px
classDef api fill:#fff3e0,stroke:#e65100,stroke-width:2px
classDef code fill:#f3e5f5,stroke:#7b1fa2,stroke-width:2px
classDef disabled stroke-dasharray: 5 5,opacity: 0.5
class n0 trigger
class n4 ai
class n3 aiModel
class n1,n7 database
class n2,n6 api
classDef customIcon fill:none,stroke:none
class n2,n6 customIcon
Why This Matters: Turning “Topics” Into Actual Posts
A topic list is not a content pipeline. The slow part is everything between “this would be a good idea” and “here’s a draft I can post without cringing.” You read a few articles, grab quotes, try to find a fresh angle, then rewrite it all in a tone that sounds like you. Do that a few times a week and you’re spending whole mornings just getting to a first draft. And honestly, the mental load is worse than the writing. You keep switching contexts, second-guessing your angle, and losing track of what’s already been covered.
It adds up fast. Here’s where it breaks down in the real world.
- You burn about an hour per post just on reading, highlighting, and pulling a coherent takeaway.
- Drafts end up inconsistent because you’re writing in bursts between meetings.
- Good ideas sit in “To do” forever, so your posting cadence becomes random.
- Finding or generating a matching image becomes a last-minute scramble, which often means you skip it.
What You’ll Build: A Topic-to-Draft LinkedIn Generator in Google Sheets
This workflow takes one topic from your Google Sheet and turns it into a complete, review-ready LinkedIn post draft with a relevant image link. You run it manually when you want a new draft (or later, you can schedule it). n8n finds the first row marked “To do,” sends that topic to Dumpling AI to pull the top three related articles, then passes those articles into a LangChain agent powered by GPT-4o. The agent summarizes the sources, writes a friendly, insightful post, and creates an image prompt that matches the angle of the draft. Finally, n8n generates the image via Dumpling AI, grabs the image URL, and writes everything back into the same Google Sheets row, flipping the status to “created.”
The workflow starts with a topic sitting in Sheets. From there, it collects real source material (three articles), then uses GPT-4o to produce a draft that sounds human and intentional. Last, it adds a usable image URL and updates your pipeline so nothing gets lost.
What You’re Building
| What Gets Automated | What You’ll Achieve |
|---|---|
|
|
Expected Results
Say you publish three LinkedIn posts per week. Manually, a typical post might take about 45 minutes to read a few sources, then another 45 minutes to write and polish, plus maybe 15 minutes to find or create an image. That’s roughly 5 hours a week for three posts. With this workflow, you spend about 5 minutes adding topics and marking them “To do,” then about 10 minutes waiting for the draft and image URL to land back in Sheets. You still review and tweak, but most teams get those three drafts ready in about an hour total.
Before You Start
- n8n instance (try n8n Cloud free)
- Self-hosting option if you prefer (Hostinger works well)
- Google Sheets for your topic queue and draft storage.
- Dumpling AI to search articles and generate images.
- OpenAI API key (get it from your OpenAI dashboard)
Skill level: Beginner. You’ll connect accounts, paste API keys, and match a few column names.
Want someone to build this for you? Talk to an automation expert (free 15-minute consultation).
Step by Step
You choose when to run it. The workflow starts with a manual trigger inside n8n, so you can generate drafts on demand while you’re setting things up.
Google Sheets provides the next topic. n8n looks for the first row where the status is “To do,” then uses that topic text as the input for everything that follows.
Dumpling AI collects real source material. The workflow sends your topic to Dumpling AI’s search endpoint, which returns the contents of the top three related articles so you’re not drafting from thin air.
GPT-4o produces the draft and image prompt. A LangChain agent analyzes the articles, writes a LinkedIn post in a friendly, insightful tone, and generates an image prompt that matches the post’s angle.
Your sheet gets updated with everything. n8n generates an image from the prompt, captures the image URL, and writes the post text, image prompt, and imageURL back into the same row, marking it “created.”
You can easily modify the prompt and the sheet columns to fit your voice and workflow. See the full implementation guide below for customization options.
Step-by-Step Implementation Guide
Step 1: Configure the Manual Trigger
Start the workflow with a manual trigger so you can test the end-to-end LinkedIn drafting flow on demand.
- Add Manual Run Trigger as the starting node.
- Keep all default settings (no configuration required).
Step 2: Connect Google Sheets
Pull one pending topic from your content library spreadsheet and later write back the generated post and image URL.
- Add Retrieve Pending Topics and set Document to your
content libraryspreadsheet and Sheet toSheet1. - In Retrieve Pending Topics, set the filter to Status equals
To doand enable Return First Match. - Credential Required: Connect your googleSheetsOAuth2Api credentials in Retrieve Pending Topics.
- Add Update Sheet with Output and set Operation to
update. - In Update Sheet with Output, map columns to values:
- Set Topic to
{{ $('Retrieve Pending Topics').item.json.Topic }}, Status tocreated, Content to{{ $('Parse Post Fields').item.json.postText }}, and Image to{{ $json.images[0].url }}. - Credential Required: Connect your googleSheetsOAuth2Api credentials in Update Sheet with Output.
Step 3: Set Up the Scraping and AI Drafting
Use Dumpling AI to scrape three articles and then generate a LinkedIn draft using GPT-4o connected to the agent.
- Add Dumpling Search Scraper and set URL to
https://app.dumplingai.com/api/v1/search. - Set Method to
POST, Send Body totrue, and Body Content Type toJSON. - Set JSON Body to
{ "query": "{{ $json.Topic }}", "numResultsToScrape": "3", "scrapeResults": "true" }. - Credential Required: Connect your httpHeaderAuth credentials in Dumpling Search Scraper.
- Add Create LinkedIn Draft and set Text to
Article 1:{{ $json.organic[0].scrapeOutput.content }} Article 2:{{ $json.organic[1].scrapeOutput.content }} Article 3:{{ $json.organic[2].scrapeOutput.content }}. - Keep Prompt Type as
defineand paste the provided system message in System Message. - Add GPT-4o Chat Model and set Model to
gpt-4o-miniwith Response Format set tojson_object. - Credential Required: Connect your openAiApi credentials in GPT-4o Chat Model. This model is connected as the language model for Create LinkedIn Draft.
json_object, Parse Post Fields will fail to parse postText and imagePrompt.Step 4: Parse the Draft and Generate the Image
Extract the JSON output from the agent and turn the image prompt into a generated image URL.
- Add Parse Post Fields and create two assignments: postText and imagePrompt.
- Set postText to
{{ JSON.parse($json["output"]).postText }}. - Set imagePrompt to
{{ JSON.parse($json["output"]).imagePrompt }}. - Add Dumpling Image Generator and set URL to
https://app.dumplingai.com/api/v1/generate-ai-image. - Set Method to
POST, Send Body totrue, and Body Content Type toJSON. - Set JSON Body to
{ "model": "FLUX.1-pro", "input": { "prompt": "{{ $json.imagePrompt }}" } }. - Credential Required: Connect your httpHeaderAuth credentials in Dumpling Image Generator.
Step 5: Test and Activate Your Workflow
Run the workflow manually to confirm each step produces the expected outputs, then activate it for production use.
- Click Execute Workflow to run Manual Run Trigger end-to-end.
- Verify Dumpling Search Scraper returns three scraped article contents in organic.
- Confirm Create LinkedIn Draft outputs a JSON string with
postTextandimagePrompt, and Parse Post Fields splits them correctly. - Check Dumpling Image Generator returns
images[0].urland Update Sheet with Output updates the row status tocreatedwith content and image URL. - When everything looks correct, toggle the workflow to Active for production use.
Troubleshooting Tips
- Google Sheets credentials can expire or need specific permissions. If things break, check the n8n Credentials area and confirm the connected Google account can edit that sheet.
- Dumpling AI endpoints can fail quietly if your API key is missing, revoked, or out of quota. Open the HTTP Request node’s last execution to see the status code and response body first.
- GPT-4o outputs depend on your prompt. If the drafts feel generic, bake in brand voice (examples, do-not-say phrases, and formatting rules) in the LangChain agent prompt so you’re not “fixing it in editing.”
Quick Answers
About 30 minutes if your accounts and sheet are ready.
No. You’ll connect Google Sheets, paste API keys, and adjust a couple of fields.
Yes. n8n has a free self-hosted option and a free trial on n8n Cloud. Cloud plans start at $20/month for higher volume. You’ll also need to factor in OpenAI and Dumpling AI usage, which usually lands around a few cents per draft depending on article length and image generation.
Two options: n8n Cloud (managed, easiest setup) or self-hosting on a VPS. For self-hosting, Hostinger VPS is affordable and handles n8n well. Self-hosting gives you unlimited executions but requires basic server management.
Yes, and it’s the best part. You can change the Google Sheets “Retrieve Pending Topics” step to pull from Airtable, Notion, or an RSS list, then keep the same GPT-4o drafting step. Common tweaks include adding your brand voice rules inside the “Create LinkedIn Draft” prompt, changing the number of sources Dumpling AI pulls, and updating which columns get written back in the “Update Sheet with Output” step.
Usually it’s expired Google authorization in n8n or the connected account doesn’t have edit access to the sheet. Reconnect the Google Sheets credential, then re-check the spreadsheet ID and tab name. If the workflow can read but not write, it’s almost always permissions. Occasionally it’s a simple mismatch too: your sheet column headers don’t match what the update step expects.
On self-hosted n8n there’s no execution limit, so it mostly depends on your server and API rate limits.
For this use case, n8n is usually the smoother choice because you’re doing multi-step AI work (search, summarization, field parsing, then image generation) and you may want to self-host when volume grows. Zapier and Make can absolutely do it, but costs can climb once you chain several actions per draft and start running it often. n8n also makes it easier to inspect raw responses when something looks off, which matters with AI outputs. If you want a simple two-step “topic in, post out,” those tools can be faster to get running. Talk to an automation expert if you’re unsure.
Once this is in place, your sheet stops being a wish list and starts behaving like a real pipeline. The workflow handles the repetitive parts, and you keep control where it matters: the final review.
Need Help Setting This Up?
Our automation experts can build and customize this workflow for your specific needs. Free 15-minute consultation—no commitment required.
