DEV Community: HaoPeng Zhang

Writing content hooks is hard — so I automated it with n8n (free template, now in the official n8n gallery)

HaoPeng Zhang — Tue, 06 Oct 2026 16:28:42 +0000

Writing the hook is the hardest 5% of a piece of content. You can have a great idea and still lose the reader in the first sentence.

I got tired of staring at a blank first line, so I built a small n8n workflow that turns a topic + audience + tone into 10 ready-to-use hooks. It just got accepted into the official n8n template gallery, so you can install it directly inside n8n.

What it does

Input: a topic, a target audience, and a tone.
Output: 10 short hooks in mixed styles, returned as JSON.

That's four nodes:

Manual Trigger
Edit Fields (Set) → your topic / audience / tone
OpenAI (GPT-4o-mini)
Output → the hooks field

The one trick that matters

Most people prompt "write me 10 good hooks." That is exactly why the output feels generic — "good" is not a spec.

Instead, the prompt writes hooks by style. Each style forces a different angle:

Contrarian — "Everything you know about X is wrong"
Curiosity gap — "Nobody talks about the part of X that actually matters"
Result-first — "I cut X from 6 hours to 20 minutes"
Story — "Three weeks ago I almost quit X"
Question — "What if X is the reason your Y isn't working?"

Ten hooks, five styles, two each. Instant variety, zero blank-page anxiety.

Sample run

Topic: AI automations for small teams · Audience: busy founders · Tone: direct

1. I replaced 3 SaaS subscriptions with one n8n workflow.
2. Nobody tells founders that the AI you actually need is boring.
3. Your team doesn't have an AI problem. It has a copy-paste problem.
4. Two weeks ago our Monday report took 4 hours. Now it's a Slack ping.
5. What if the automation you keep avoiding is a 20-minute build?

Install it

Official n8n template: https://n8n.io/workflows/20418-generate-social-content-hooks-with-openai-gpt-4o-mini/
GitHub (MIT, importable JSON): https://github.com/zhp910318/n8n-ai-automation-lab
Free download plus more templates at my store: https://payhip.com/haopengailab

Steps: open the template → connect your OpenAI credential → set topic / audience / tone → run.

Bonus: what I learned getting a template approved

If you publish to the n8n gallery, two things trip people up:

Sticky notes are reviewed by hand. Use section headers like ## 1. Trigger and input, give them a color, and size them to cover the nodes they describe.
The Setup section needs real length (50+ words). Write it for a stranger, not for yourself.

The review runs AI pass → human pass. Mine came back once with notes, I fixed them, and it went live.

If you would rather not build it

The free workflow handles hooks. The paid ones handle the rest of the pipeline — content repurposing, inbox triage, and SEO blog generation — and are all on the store. If you want any of it wired into your own stack, I do fixed-price done-for-you builds.

I automated my inbox triage with n8n + GPT-4o-mini (copy the exact workflow)

HaoPeng Zhang — Tue, 06 Oct 2026 00:06:45 +0000

Inbox zero is a lie. Triage is the real skill.

You don't need an empty inbox. You need to know which 12 of your 200 unread emails actually need you today — and which ones can die quietly.

I got tired of scanning my inbox every morning, so I built a small n8n + GPT-4o-mini workflow that reads new mail, sorts every message into one of four buckets, and drafts a reply for the ones that need one. It runs on a schedule, costs pennies, and takes about 15 minutes to set up.

Here's the whole thing — plus the one design decision that made it actually usable.

The idea: force a fixed taxonomy

The first version I built asked the model "is this email important?" It was useless. Every email came back "somewhat important."

The fix was to stop asking for a judgement and start asking for a single label from a fixed taxonomy. Four buckets, no hedge, no score:

Reply today — a real human is waiting on a decision or an answer.
Task, not today — actionable, but it belongs in a to-do list, not your face right now.
Read later — newsletters, FYIs, things worth a skim.
Noise — promos, cold outreach, notifications you'll never open.

One label per email. That's it.

The prompt (this is the whole trick)

You are an inbox triage assistant. Read the email below and return exactly one of these labels on the first line: REPLY_TODAY, TASK_LATER, READ_LATER, NOISE.

Rules:

Only use REPLY_TODAY if a specific person is blocked on your answer.

Marketing, cold sales, and automated notifications are always NOISE.

If you choose REPLY_TODAY or TASK_LATER, write a one-sentence draft reply (max 40 words) on the second line, in the sender's language, in a friendly professional tone.

Subject: {{ $json.subject }}
From: {{ $json.from }}

Body:
{{ $json.snippet }}

Two things matter here: the label set is closed, and the draft is only generated for the two buckets that need one. That keeps token usage (and cost) low.

The 4 nodes

Schedule Trigger — every weekday at 08:00.
Gmail → Get Many Messages — filter in:inbox is:unread newer_than:1d.
OpenAI (GPT-4o-mini) — the prompt above, text mode, temperature 0.
Router → Gmail / Sheets — REPLY_TODAY gets starred + draft attached; TASK_LATER gets appended to a Google Sheet; the rest get labeled and archived.

That's a real workflow, not a toy. The router is just a Switch node on the first line of the model's output.

3 rules that made it usable

Temperature 0. Classification should be boring and repeatable, not creative.
Send the snippet, not the full body. The model doesn't need a 40-line signature block to know an email is a newsletter — and it halves your token bill.
Never auto-send. The workflow drafts. You still tap send. That single decision is the difference between a helpful assistant and a horror story.

Want it without rebuilding it?

Rebuilding this from scratch (Gmail OAuth, the filter, the parser, the router) is the boring part. I packaged the polished version as a ready-to-import n8n template:

AI Inbox Triage & Reply Drafter — the workflow above, fully wired, with the router and the Sheets log: https://payhip.com/b/W0mLE ($24)
Want a taste first? My free AI Hook Generator workflow (turn any topic into scroll-stopping hooks) is free to download: https://payhip.com/b/ulER8
All the free templates + copy-paste JSON live in this MIT repo: https://github.com/zhp910318/n8n-ai-automation-lab
More workflows: https://payhip.com/haopengailab

Triage your inbox once, and you'll never go back to scanning.

I built a free n8n workflow that writes scroll-stopping content hooks

HaoPeng Zhang — Mon, 05 Oct 2026 16:06:39 +0000

Hook writing is the highest-leverage 15 minutes in content. If the first line doesn't stop the scroll, nothing else gets read. I kept rewriting the same hook prompts by hand, so I automated it with n8n + OpenAI (gpt-4o-mini).

Here's the workflow, why it's structured this way, and how to run it in 2 minutes.

What it does

You give it a topic, audience, and tone. It returns 10 hooks across 5 proven styles:

Curiosity gap
Contrarian take
Specific result / number
Mistake / warning
Story opening

Why a workflow and not just "ask ChatGPT"

Repeatable — same input gives the same format every time, so you stop re-prompting.
Batchable — run a whole week of posts in one execution.
Composable — the output feeds a scheduler, a sheet, or your CMS.

The 4 nodes

Form Trigger — collects topic, audience, tone.
OpenAI (gpt-4o-mini) — a system prompt that enforces 5 styles x 2 variants.
Code node — splits the output and tags each hook with its style.
Output — a clean list you can copy or push downstream.

The one prompt trick that matters

Ask for styles, not "good hooks". "Write 10 hooks" gives you 10 flavours of the same sentence. Specifying 5 distinct styles forces range. The core system prompt:

You are a direct-response copywriter. For the topic and audience given,
write 10 hooks: 2 each in these styles — curiosity gap, contrarian,
specific result/number, mistake/warning, story opening.
Each hook <= 12 words. No hashtags. Use emojis only if tone=playful.
Return a numbered list with the style label in brackets.

That single constraint is the difference between a generic list and something you can actually post.

Try it

The workflow is free and MIT-licensed — same engine I use before hooks go into a repurposing pipeline.

Free n8n workflow (JSON + README): https://payhip.com/b/ulER8
Repo with copy-paste examples: https://github.com/zhp910318/n8n-ai-automation-lab

If you want the rest of the pipeline — turning one piece into a week of posts, AI inbox triage, or an SEO blog engine — they live here: https://payhip.com/haopengailab

Building this in public. If the setup notes or sticky notes are unclear, tell me and I'll fix them.