And the thing nobody tells you. Also no charge.
If you've watched someone get better output from the same tools you use, this is usually the difference:
They run the whole job in one chat. You start a new one for every prompt.
When every step happens in the same conversation, the model still knows your audience, your angle and the promise you settled on six prompts ago. By the time it drafts, it's working with everything you told it. Start fresh each time and it's guessing from one paragraph, every time.
Same prompts. Different results. The difference is which window you're typing in.
That one applies to the five free prompts too. Use it today without buying anything.
One scroll down, the paid one
By step four, the piece has lost the shape you gave it in step one.
The five prompts above are stage one. They find the thing and split it. What they don't do is carry a decision forward, and that's where the time actually goes. Content Workflow is a seven-stage workflow for developing, judging, publishing and improving one real content idea.
$39 · one payment · works in ChatGPT or Claude
The problem isn't that you can't produce. It's that you can't judge.
You've got no shortage of ways to make something. What you haven't got is a standard to hold it against, so the decision to publish comes down to whether you're sick of looking at it.
Everyone sells prompts that write. Almost nobody sells the one that checks.
The scorer, in full
Paste anything you've published and it scores seven things, nought to three each:
- Specificity
- Does every claim carry a number, a name, a situation or a quote. Then it quotes your vaguest sentence back at you.
- Rhythm
- Variation, read against the format you chose. Sentence length in a text post. Progression across slides in a carousel. Spoken pacing and beat changes in a video.
- Machine tells
- Em dashes, "it's worth noting", "delve", "in today's landscape", the not-X-but-Y tic. Listed by name, wherever they appear.
- Stance
- Could anyone have written this, or would a named group actually push back. It tells you who, and why.
- Earned claims
- Every claim sorted into evidenced, true but unshown, opinion, or not yours to make. Anything reading like a result claim or an implied income promise gets flagged and rewritten.
- Opening
- Does it start at the real start. For a video, does frame one work with the sound off and no context.
- Reader payoff
- Is there one specific thing they can do today. It names it, in their words.
What comes back: a table of the seven with one line of evidence each, the total out of 21, the three fixes that move it most, and the single worst sentence in your piece quoted with a rewrite.
Underneath that, unscored, a voice read-out: how far the piece drifted from your own writing, which three sentences drifted furthest, and the corrected version of each.
And a rule it won't break:
If specificity, earned claims or reader payoff scored 0, tell me the piece isn't ready and stop there. Don't offer a polish.
I ran this on a post of mine that was already live and got 11 out of 21.
The scale is strict on purpose. A 3 is rare. Single digits on a first pass are normal. Mine was 11.
Here's the part I'd rather tell you. When I first built that scorer, it was broken. It counted five dimensions out of seven and printed a total out of 15 as though that were the whole thing. The one it silently dropped was the check that flags unearned claims. I only found it by running it on my own writing.
Seven stages. Each one feeds the next.
Exactly what is inside, section by section.
- 01Teach it your voiceYou paste in things you've already written. It returns a Writing Spec: how you open, how you close, what you never say, what you do instead of hedging. Then it writes with the Spec, criticises its own output, and fixes the parts of the Spec that were too vague to work from.
- 02Find the angleWhat your audience is quietly tense about, what everyone in your space is already saying, and the posts hiding in your last two weeks of actual work. Ends with one angle scored against five criteria, not twenty ideas.
- 03Lock the pieceThe promise, the claim, the evidence, and the platform and format, chosen separately and for a reason. Then a skeleton built to that format rather than to a generic outline.
- 04Write itNative to the platform and format from the first sentence. No neutral draft converted afterwards. Then a tightening pass of up to 30 percent that tells you what it deliberately refused to cut, and the ending, chosen from four with the cost of each.
- 05Score itThe scorer above, a cold read of your opening by someone who's never heard of you, and the boring pre-publish list that catches the broken link and the number that doesn't match the one three lines up.
- 06Repurpose itIt pulls the finished piece apart into its claim, evidence, story, instruction, contrarian element and question, then finds six different jobs those parts can do. It ranks them, tells you which three are worth building, and builds them to production spec. Job first, then format.
- 07Learn from itWhat to keep, what to stop, what to test next, and the next three decisions. It works out whether you have enough data for a real read, and says so plainly when you don't rather than inventing a conclusion.
Every prompt tells you when a working CMO actually runs it.
I'm a working CMO, fourteen years in marketing. Different industry to yours, same job underneath: getting a stranger to trust a screen enough to hand over money before they've seen the thing.
So the prompts carry a line saying where they sit in a real week. What runs on a Monday morning. What happens before a landing page goes live. What I do after something publishes.
That line is the part you can't get from a free prompt, and it isn't something anyone can write without the day job.
What you actually get
- An interactive page, not a PDF. Every bracket in every prompt is a box. Type your audience once and it drops into every prompt that needs it.
- One connected working context. Your Writing Spec, your audience, what you sell, your locked brief, your platform and format. Filled in once, carried through all seven stages.
- Construction rules that assemble themselves. Pick a platform and a format and the right rules drop into every prompt that needs them, including the production layer for video and the claims guard for what that combination is not allowed to say.
- The prompt rewrites itself as you type, with your words highlighted, and the copy button gives you the filled version.
- Nothing you type leaves your browser. It saves on your own device so you don't lose an hour's work, and it goes nowhere else.
- Download it as one file. Works offline, yours to keep, keeps working if the link ever changes.
What it doesn't do
It won't plan your month. It takes one strong argument and turns it into selected derivatives built for different jobs and formats. It won't hand you thirty different arguments.
It won't build you a product. If you're looking for something that takes you from nothing to a thing you can sell tonight, this isn't it.
It won't write in your voice by magic. Stage 1 builds a spec from your actual writing, and the rest of the workflow follows it. Feed it three thin samples and you'll get a thin spec, and it will tell you so rather than pretend.
Before you buy
All sales final, and here's what that does and doesn't mean.
There is nothing to return. The moment you pay you have everything. At checkout you confirm that you understand the product is delivered instantly and is non-refundable.
Rights you have as a consumer that cannot be waived are unaffected.
Final sale is not a licence to leave you stuck. If any of the following happens, email hello@onlinehub.marketing and I will sort it out:
- You cannot access what you paid for. I will get you access
- You were charged twice, or charged for something you did not buy. I will refund it
- You bought the wrong product. Tell me before you use it and I will swap it
These are not really refunds, they are me delivering what you paid for. That obligation does not expire.
Not sure it fits? Email me first at hello@onlinehub.marketing. If it is wrong for you I will tell you.
Purchases are processed by Polar as merchant of record, which means Polar is the seller and handles payment, invoicing and tax. Full terms on the refund policy page.
Questions
- What's the difference between the free prompts and the paid one?
- The five free prompts take one idea and split it six ways. That's part of stage two of seven. Content Workflow adds the voice spec, the brief, the drafting, the scorer, the repurposing engine and the review, and carries your context through all of it.
- How many prompts is it?
- The workflow includes 20 purpose-built prompts, but you will not use every prompt for every piece. Stage 6 runs once per derivative, and Stage 7 waits until you have something worth reading.
- Does it work in ChatGPT or Claude?
- Both, and anything else that holds a conversation. Run the whole workflow in one chat, one chat per piece.
- What platforms and formats does it cover?
- LinkedIn and Instagram, as text posts, carousels, talking-head video, and Stories. Platform and format are chosen separately, so a LinkedIn carousel and an Instagram carousel get different rules.
- Do I need a content history to start?
- No. Stage 1 needs three to five things you've already written, any channel. Stage 7 has a path for people who don't have numbers yet, and it will tell you what it can and can't conclude from what you have.
- How long does a run take?
- The scorer on its own is about ten minutes. A full run the first time is about an hour, most of it thinking rather than typing.
- Is there a refund?
- No, and the section above says exactly what that does and doesn't mean. If it's wrong for you, email me before you buy and I'll say so.
$39. One payment.
Seven stages for getting one idea published properly, and the one prompt that tells you whether it's ready.
$39 · one payment · instant access