Voice OS is an AI writing system that writes LinkedIn posts in your own voice. It is built from a long, structured interview with you and holds four things: a voice guide that describes how you actually talk, a bank of your real stories, a bank of your opinions, and your positioning. You load it into ChatGPT or Claude, and the drafts come out sounding like you, with your facts, instead of like every other AI post on the feed.
What is Voice OS?
Most AI-written posts sound the same because the AI has nothing to go on but the prompt. Ask it for a post about leadership and it gives you the average of every post about leadership it has ever read. Smooth, confident and empty.
Voice OS fixes the input, not the tool. It is a set of documents about you, written in a form an AI model can follow: how you build a sentence, which words you use and which you never would, the stories only you can tell and the opinions you will actually defend. Put that in front of ChatGPT or Claude and the output changes completely, because the model is no longer guessing.
It is for founders, experts and business owners who want to post on LinkedIn in their own voice, regularly, without hiring a ghostwriter for every post.
Why does AI writing sound generic?
Three reasons, and each one has a fix.
- No real material. Without your stories, the model invents them or stays vague. Vague posts do not get remembered. The fix is a story bank: real events from your work, with the names, numbers and details that make them yours.
- No point of view. A model tries to be balanced. People follow people who think something. The fix is an opinion bank: the views you hold, why you hold them, and how far you are willing to push each one in public.
- No voice. Models drift to the same rhythm and the same stock phrases, the ones readers now spot in a second. The fix is a voice guide built from how you speak, including a list of what you never say.
None of this needs a better prompt. It needs better material about you, written down once and reused every time.
What is inside a Voice OS?
A voice guide
How you actually talk, taken from recordings of you talking, not from how you think you write. Sentence length and rhythm. The phrases you repeat. How formal you are, how much humour you use, how direct you are, each on a simple scale. How you open a point and how you close one. And a list of anti-patterns: the things that would make a post sound like somebody else. It ends with a short test any draft has to pass before you post it.
A story bank
Thirty or more scenes from your working life, each written down with the details that matter: who, where, how much, what was said, what you learned. These are the raw material for posts. A good story bank means the model never has to make anything up, and you never run out of things to say. My own Voice OS holds 34 stories.
An opinion bank
Twenty-five or more views you hold about your work and your industry, each with the reason behind it and a note on how risky it is to say in public. Some are safe. Some will start an argument. You decide which ones go out. Mine holds 31.
Your positioning
Who you are writing for, what they struggle with, what you stand against, and the few themes you want to be known for. This keeps every post pointed at the right reader, so the people who follow you are the people who could one day buy from you.
Set up where you already work
Everything is loaded into a project in ChatGPT or Claude, with instructions that tell the model how to use each document. You open it, give it a topic or a story number, and get a draft in your voice. You edit, you post.
How is it built?
It starts with a long, structured conversation. I ask about your work, your clients, the jobs that went wrong and the ones you are proud of, the things your industry gets wrong, and what you would say if nobody could stop you. It is recorded, because the way you talk is the evidence.
From that conversation I write the voice guide, the story bank, the opinion bank and the positioning, then set up your workspace and test it. If a draft does not sound like you, the guide changes until it does. You get the documents to keep. They are yours, and they work with whichever AI tool you prefer.
Want a quick taste first? The free voice check analyses a sample of your writing and shows how your voice comes across. It takes about two minutes.
What does it look like day to day?
You open your Voice OS project and ask for a post. For example: "Story 12, opinion 4, for the people who keep putting off their website." The model picks up the story, the opinion and the reader, and writes in your rhythm with your words. You read it, fix the line that is not quite right and post it.
A good rhythm is two or three posts a week. The time goes into reading and adjusting, not into staring at a blank page. And because every fact comes from your own banks, you are never posting something that did not happen.
The banks grow with you. When something worth telling happens, you add it as a new story, and it is there for the next post.
Voice OS or the full Forge Lead service?
They are for different people.
- Voice OS is for anyone who wants to run their own LinkedIn and write in their own voice: founders, consultants, coaches, experts and business owners. You stay in control and do the posting yourself.
- The full service is for UK bespoke makers who sell through architects. I write the posts, connect you with the right architects and start the conversations, and you take the calls. See how that works.
If you are a maker who would rather run LinkedIn yourself, Voice OS is where to start.
Who is it not for?
- People who want somebody else to do all of it. Voice OS writes the drafts. You still read, decide and post.
- People with nothing to say yet. It works from real stories and real opinions. If the work has not happened yet, there is nothing to put in the banks.
- People chasing likes. It is built to make you known by the right readers, not to game the feed.
Questions about Voice OS
Does Voice OS work with ChatGPT and Claude?
Yes. It is a set of documents and instructions, so it works with either. Most people already pay for one of them, and that is all you need.
Will people know the posts were written with AI?
The drafts come from your own stories and follow your own way of speaking, and you edit every one before it goes out. The aim is a post you would have written yourself if you had the time. The voice guide includes a list of the patterns that give AI writing away, so the model avoids them.
How long does it take to build?
It starts with one long conversation with you. The documents and the workspace follow from that. On the call we agree the dates that suit you.
Do I need to be good at writing?
No. The voice guide is built from how you talk, not how you write. Many people who find writing hard talk very well, and that is what the system captures.
Can I use it in another language?
Yes. The voice guide is written for the language you post in. My own is in Polish.
What happens if my business changes?
You update the positioning and add new stories and opinions. The documents are yours, so you can change them whenever you like, or ask me to.
Want posts that sound like you?
Ask about Voice OS on a short call, or send a note.
Read more: LinkedIn for bespoke makers · About Peter Solarski · All questions