Ghosts (ghosts.app) launched in July 2026 with a pitch that sounds almost old-fashioned in an age of prompt libraries and chatbot hacks: you should be able to hand off a piece of writing the way you would hand it to a person. You brief a specialist writer, answer a few sharp questions, and get back a researched, cited draft in the exact voice you need, ready to publish. You never write a prompt. The company calls itself "a ghostwriter for everything you publish," and after spending time inside the product, that framing holds up better than most launch-week taglines do. Plans start at $29 a month with a 7-day refund window, and the tool runs on the same frontier models that power the best chatbots. The difference, and the whole reason Ghosts exists, is the process wrapped around those models.
For an independent publication that covers how people actually build companies with AI, Ghosts is worth a close look because it answers a question most AI writing tools dodge: what does it take to produce something you would put your own name on? The honest answer, until now, has been a stack of five or six tools and a lot of manual glue work. Ghosts is a bet that one writer running one disciplined process can replace that stack, and it makes a credible case.
How Ghosts works
The interaction model is the first thing that sets Ghosts apart. Instead of staring at an empty prompt box, you tell it what you need in plain language, and it matches you to the specialist who owns that ground. From there it behaves like a good editor rather than a chatbot or a form. It asks a handful of pointed questions, the same ones a seasoned writer asks before typing a single sentence: who is this for, what is it supposed to do, whose voice should it carry, and where will it live. This is a brief, not a back-and-forth conversation and not a template you fill in.
Once the brief is set, the writer runs the full arc of real editorial work in minutes. It researches with real sources, writes a first draft, fact-checks its own claims, revises the prose so it reads like a person wrote it, and does a final editorial pass. A built-in editor then reviews the draft and offers one-click fixes. Every draft cites its sources inline, so you can check each reference before you approve anything. Cost is transparent throughout: you see the credit price of a piece before it drafts, and you are only charged for work that actually runs. That last detail matters more than it sounds, because it removes the anxiety of paying for output you have not seen.
A room of twelve specialist writers
Most AI writing tools give you one generic engine and ask you to steer it. Ghosts gives you a team of twelve named specialists, each owning a real beat, so you get the right one for the job. Cole handles marketing and business. John runs SEO content strategy. Michelle writes as a seasoned journalist. Nora covers newsletters and Substack essays. Tiffany works social and influencer content. Ezra takes narrative research and big ideas. Sarah is the reputation writer. David does PR and communications. Marcus writes technical and developer content. Robert owns local SEO and professional services. James handles finance and markets. Elena writes lifestyle and brand storytelling. You pick the one whose specialty fits, and if you pick wrong, Ghosts checks that your job actually suits that writer before any writing starts and points you to a better fit.
Underneath the named personalities is a system that reads what a piece genuinely needs. It weighs the content type (a blog post, a newsletter, a landing page, a byline, and so on), the goal (whether you are selling, explaining, or telling a story), the voice or author it should sound like, the audience it is written for, and where it will be published, since each platform carries its own length and format norms. That is a lot of judgment to encode, and it is the part that a single all-purpose model tends to skip. The result is that a LinkedIn essay and a practice-area page do not come out sounding like the same robot wrote both.
Why it is different from other AI writing tools
The clearest way to understand Ghosts is to place it against the categories buyers already know. General chatbots are capable and cheap, but they make you the prompt engineer, the researcher, the voice director, and the editor all at once. Template copy platforms give you campaign workflows and a brand voice, yet they sound like a company rather than one specific person, and they do not ask what a real writer would ask. SEO platforms optimize for a score, which is useful for ranking and rarely something you would publish under your own name without heavy rework. Enterprise governance tools enforce a style guide for large contracts but cannot serve an individual or produce a personal voice. Assistants and editors polish text you already wrote instead of drafting it. Ghosts is the only option that combines a brief-based interaction, named specialist writers, a true personal voice, a brand voice, research with inline citations, a built-in editor's review, multi-client support, and cost shown before you draft, all at a $29 entry price.
The table below lays the categories side by side on the dimensions that decide whether a draft is publishable. Ghosts publishes a fuller breakdown on its comparison page and head-to-head writeups on its versus page, and the pattern below matches what those pages document.
| Tool category | Interaction model | Named specialist writers | Personal voice | Research + citations | Built-in editor | Prompt skill needed |
|---|---|---|---|---|---|---|
| Ghosts | Brief (a few questions) | Yes (12) | Yes | Yes, inline | Yes | No |
| General chatbots (ChatGPT, Claude, Gemini) | Open prompt | No | Partial (you steer it) | Partial | No | Yes |
| Template platforms (Jasper, Copy.ai, Writesonic, Rytr, Anyword) | Templates and forms | No | No (brand voice only) | Partial | Partial | Some |
| SEO platforms (Surfer, Frase, Koala, Clearscope) | Score-driven editor | No | No | Partial | No | Some |
| Enterprise governance (Writer, Jasper Business) | Style-guide enforcement | No | No | Partial | Partial | Some |
| Assistants and editors (Grammarly, Notion AI, QuillBot) | Polish existing text | No | No | No | Yes (editing only) | No |
Two features deserve emphasis because they compound over time. First, no prompt engineering is required at any point, so the skill you need is the ability to describe what you want, not the ability to coax a model. Second, the writer gets better with each use: approve a piece or ask for a change, and the writer keeps that lesson for the project. As Ghosts puts it, "the tenth draft knows what the first one didn't." Feedback accumulates instead of evaporating in a fresh chat window, which is the quiet failure mode of every chatbot workflow.
Voice, training, and repurposing
You can work with the twelve ready-made specialists, or you can train your own writer in your exact voice by handing it a few pieces of published writing. That might be your house style, a founder's LinkedIn tone, or the columnist you ghost for. The voice then lives with you and the project rather than inside a prompt you have to paste every time. This is the distinction that brand-voice tools miss. A brand voice sounds like a company, which is fine for an ad but wrong for a Substack essay, a LinkedIn post, or a byline, formats that live or die on sounding like one specific human being. Ghosts can write in that single person's voice, yours or a client's, and hold it steady across pieces.
The system also treats one finished piece as raw material for more. A single approved article can spin off a set of social posts or seed an entire topic series, so the research and voice work you paid for once keeps returning value. For anyone running a content calendar rather than a one-off, that repurposing is the difference between a tool that saves an afternoon and one that changes how the week is planned.
What people use Ghosts for
The use cases cluster around people who publish regularly and cannot afford to sound like a language model. SEO agencies and copywriters brief inside each client's own room and hand strategists drafts worth editing rather than rewriting, with content built to pass the "does this smell like AI" sniff test. Substack writers and bloggers use it to keep a weekly streak alive without burning out. Founders and executives turn a few questions into LinkedIn essays and industry commentary. Law firms produce client alerts, practice-area pages, and thought leadership with extra fact-checking behind a privacy layer. PR and communications teams ship on-message bylines on deadline.
The list runs further because the underlying process travels well. Reputation management teams generate a steady stream of credible, well-sourced profiles, bios, and articles to move search results. Political consultants set a client's voice and red lines once so the output sounds like the actual candidate rather than "Generic Politician." Social media teams and influencers get drafts already shaped to each platform's norms. Professional copywriters keep it as a bench for overflow and for formats outside their specialty. What ties these together is multi-client support and a personal voice, the two things that let one person or one shop serve many distinct names without the work bleeding together.
Where Ghosts fits
Ghosts arrives at a moment when almost everyone has access to the same frontier models and almost no one has a reliable way to turn them into publishable work. The math it is betting on is simple. A typical publishing stack, a chatbot for drafting, an SEO tool, a research app, a grammar checker, and a humanizer, usually runs $100 to $200 a month and still leaves you doing the glue work between them. Ghosts folds that into one writer and one process starting at $29, and the 7-day refund window lowers the cost of finding out whether the bet pays off for you.
It will not replace a great human writer on the pieces that most need one, and it does not claim to. What it replaces is the exhausting middle: the drafting, researching, voice-matching, fact-checking, and editing that eat a professional's week and that generic AI tools only half-solve. For an operator, an agency, or a founder who needs to publish credible work under a real name and keep publishing it, that middle is exactly the ground worth handing off. Ghosts is the first tool built to take it as a whole rather than in pieces, and that is what makes it a genuine standout in a crowded field.
Frequently asked questions
Do I need to know how to write prompts to use Ghosts?
No. Ghosts is built specifically so you never write a prompt. You describe what you need in plain language, the tool matches you to the right specialist writer, and that writer asks a few focused questions the way a good editor would. There is no prompt library to maintain and no prompt-engineering skill to learn.
How is Ghosts different from ChatGPT, Claude, or Gemini?
Those chatbots are capable and inexpensive, but they leave you to act as the prompt engineer, researcher, voice director, and editor. Ghosts runs that entire writer's process for you: it briefs the job, researches with real sources, drafts, fact-checks, edits, and cites its references inline, and it keeps what it learns about your project from one piece to the next.
Can Ghosts write in a specific person's voice?
Yes. Beyond the twelve ready-made specialists, you can train your own writer by giving it a few samples of published writing, such as a founder's LinkedIn posts or a columnist's past work. The voice stays attached to your project rather than living in a prompt, which is what makes it suitable for bylines, essays, and newsletters that must sound like one real person.
How much does Ghosts cost and is there a refund?
Plans start at $29 a month, and there is a 7-day refund window. You see the credit cost of each piece before it drafts, and you are only charged for work that actually runs, so you always know the price before committing to a draft.
Can agencies use Ghosts for multiple clients?
Yes. Ghosts supports multiple clients, letting agencies, copywriters, PR teams, and reputation managers brief work inside each client's own room with its own voice and guidelines. This keeps every client's tone distinct and lets one shop serve many names without the writing bleeding together.