How to Automate Social Media Content with n8n + Ollama (Free Workflow Included)

Published March 24, 2026 · 8 min read · Get the free workflow →

Creating social media content for multiple platforms is time-consuming. Each platform has its own tone, character limits, and audience expectations. What works on LinkedIn doesn't work on Twitter, and neither works on Reddit.

In this tutorial, you'll build an n8n workflow that takes a single topic and generates optimized content for 4 platforms — Twitter/X, LinkedIn, Reddit, and Instagram — using Ollama for local AI processing. No API keys, no monthly costs, no data leaving your machine.

            What you'll build: A one-click workflow that generates platform-specific social media posts with AI quality review. Total setup time: ~10 minutes.
        

Why Local AI for Content Generation?

Using ChatGPT or Claude APIs for social media content works, but has drawbacks:

Factor	Cloud AI APIs	Ollama (Local)
Cost per post	$0.01–0.05	$0 (free forever)
Monthly cost (daily use)	$10–50+	$0
Data privacy	Sent to cloud	Stays on your machine
Rate limits	Yes	No
Works offline	No	Yes
Setup complexity	API key + billing	One install command

For content generation, local AI models like llama3.2 or mistral produce quality that's comparable to cloud APIs — especially for short-form social media content.

Prerequisites

n8n — Self-hosted (docker run -it --rm -p 5678:5678 n8nio/n8n) or n8n.io cloud
Ollama — Install from ollama.ai, then ollama pull llama3.2

That's it. No API keys, no accounts, no billing setup.

How the Workflow Works

You enter a topic (e.g., "benefits of self-hosted AI")
    │
    ▼
n8n sends topic to Ollama with platform-specific prompts
    │
    ├── Twitter/X: Concise, punchy, hashtags, 280 chars
    ├── LinkedIn: Professional, insights, 1200 chars
    ├── Reddit: Conversational, detailed, community-friendly
    └── Instagram: Visual-focused, emoji-rich, 30 hashtags
    │
    ▼
AI Quality Review pass (checks tone, engagement, accuracy)
    │
    ▼
4 ready-to-post pieces of content

The workflow uses two AI passes: one for generation, one for quality review. The review pass catches issues like:

Content that's too generic or salesy
Platform tone mismatches (e.g., LinkedIn post that reads like a tweet)
Missing or excessive hashtags
Grammar and clarity issues

Step-by-Step Setup

STEP 1: Download the Workflow

Download the JSON workflow file from GitHub:

# Direct download
curl -O https://github.com/bonskari/n8n-ollama-social-content/raw/main/workflow.json

# Or clone the repo
git clone https://github.com/bonskari/n8n-ollama-social-content.git

STEP 2: Import into n8n

Open n8n (usually at http://localhost:5678)
Click Add workflow → Import from file
Select the downloaded workflow.json

STEP 3: Configure Ollama Connection

In the workflow, each HTTP Request node points to Ollama's API. By default, it's configured for:

URL: http://localhost:11434/api/generate
Model: llama3.2

If your Ollama is on a different host or you want a different model, update the URL and model name in the "Set Parameters" node at the top of the workflow.

STEP 4: Run It

Click Execute Workflow. Enter your topic in the manual trigger dialog. In about 30–60 seconds, you'll have 4 platform-optimized posts ready to copy and paste.

Example Output

Input topic: "Why developers should try self-hosted AI tools"

Twitter/X Output

🔓 Stop paying for AI APIs when you can run the same models locally.

Self-hosted AI tools give you:
✅ Zero API costs
✅ Full data privacy
✅ No rate limits
✅ Works offline

Try Ollama + n8n — free, open-source, and yours to keep.

#SelfHosted #AI #OpenSource #n8n #Ollama

LinkedIn Output

I've been experimenting with running AI models locally instead of using
cloud APIs, and the results have surprised me.

Here's what changed:

→ Monthly AI costs dropped from $47 to $0
→ Sensitive data never leaves my network
→ No more rate limit errors at 2am
→ Everything works even without internet

The tools that made this possible: Ollama for running models locally,
and n8n for building automation workflows around them.

If you're building AI-powered automation and cost or privacy matters
to you, self-hosted is worth exploring.

What's your experience with local AI models?

Customization Tips

Change the model: Swap llama3.2 for mistral, gemma2, or any model Ollama supports
Add brand voice: Edit the system prompts to include your brand guidelines, tone, and key messaging
Add more platforms: Duplicate a generation node and customize the prompt for YouTube Shorts, TikTok, newsletters, etc.
Schedule it: Replace the manual trigger with a Cron node to generate content on a schedule
Connect to posting tools: Add nodes to auto-post via Buffer, Hootsuite, or platform APIs

Model	Size	Best For	Speed
llama3.2	3B	Fast, good quality general content	Fast
llama3.1:8b	8B	Higher quality, still fast enough	Medium
mistral	7B	Strong writing quality, European language support	Medium
gemma2:9b	9B	Google's model, good for factual content	Medium

More Free Workflows

Try these other free n8n + Ollama workflows:

AI Blog Writer Pipeline — Research, outline, draft, and edit blog posts with local AI
AI Email Auto-Responder — Classify emails, filter spam, draft replies automatically
All Free Samples — 3 free workflows + documentation

Troubleshooting

Ollama connection refused

Make sure Ollama is running (ollama serve) and accessible at http://localhost:11434. If n8n runs in Docker, use http://host.docker.internal:11434 instead of localhost.

Slow generation

Try a smaller model (llama3.2 instead of llama3.1:8b). If you have a GPU, make sure Ollama is using it — check with ollama ps.

Output quality issues

Upgrade to a larger model, or edit the system prompts in the workflow to be more specific about your desired output format and tone.