Ghosts

Description

Ghosts is an AI writing and optimization engine that finds and fills a site's content gaps. Pointed at a URL, it works out what the business sells and where, then lists the searches customers would type that no page there answers, including searches with no page at all and ones diluted into a page built for something else. It drafts one page per gap, shaped to a structure and citation floor measured from pages already ranking in that industry.

Rather than starting from an empty prompt box, Ghosts assembles the assignment first: a planning agent researches the topic and commits to a thesis and structure, then the draft is written against that plan. It splits writing across specialized agents, including a voice persona trained on your writing samples, an independent fact-checker that verifies every claim against the actual text of fetched sources, and editorial passes that strip machine-writing patterns. Verified numbers and dates are locked so later edits cannot quietly change them, and each draft carries a visible log of what was flagged and changed.

The voice persona learns from every edit a team makes, with filters so one odd edit does not corrupt the voice, and the learning ledger is visible and reversible. Ghosts applies E-E-A-T signals, cited sources, schema, internal links, and answer-engine formatting while drafting, and holds distinct voices for multiple clients. A dedicated law firm lane applies jurisdiction advertising rule packs while drafting, flags language against the rule, and requires a named reviewer to sign off.

Ghosts uses a credit system where drafting, researching, fact-checking, and revising all consume credits, with the cost shown before each piece. Plans include Individual at $29/month with a 7-day free trial, plus higher tiers for creators, agencies (priced per seat), and law firms. Content is never used to train AI models, workspaces are isolated with per-tenant security, and data is encrypted in transit and at rest.

Ghosts's Core Features

• Site scan that surfaces content gaps with no page of their own

• Drafts shaped to a structure and citation floor measured from ranking pages

• Multi-agent pipeline: planner, voice writer, fact-checker, and line editor

• Independent fact-checking against the real text of fetched sources

• Locked verified numbers and dates with a visible change log

• Voice personas trained on your samples that learn from every edit

• E-E-A-T signals, citations, schema, and answer-engine formatting applied while drafting

• Law firm lane with jurisdiction rule packs and named reviewer sign-off

How to use Ghosts?

1. 1

   Scan a site: Point Ghosts at a client URL to get a ranked list of content gaps, with no card required.

2. 2

   Set up a project: Define the brand, audience, and voice, and pick or train a writer on real samples.

3. 3

   Assign a piece: Choose where it will publish and give a topic, confirming anything ambiguous.

4. 4

   Let it draft: The agents research, draft to the plan, fact-check, and humanize, showing their work.

5. 5

   Review and publish: Read the editor's notes, apply the changes worth applying, and publish.

Ghosts's Use Cases

• Content gap analysis
• Agency client work
• Fact-critical content
• Law firm publishing
• Solo creator publishing

FAQ from Ghosts

• What does Ghosts do?

  Ghosts scans a site to find the searches with no page of their own, then drafts those pages to a structure measured from pages already ranking, fact-checks them against sources found in research, and handles on-page optimization while writing.

• How is Ghosts different from a single AI writing app?

  Instead of one model doing everything, Ghosts splits the work across specialized agents: a planner that researches and commits to a structure, a voice persona trained on your writing, an independent fact-checker, and editorial passes that strip machine-writing patterns.

• Does Ghosts make things up?

  Every factual claim is checked against the actual text of fetched sources rather than the model's memory. Verified numbers and dates are locked so later edits cannot change them, and the system confirms ambiguous subjects with you before drafting.

• Will it actually sound like me or my client?

  The persona is trained on writing samples, learning sentence rhythm and diction, and every edit teaches it a durable lesson. The learning ledger is visible and reversible, and the tenth piece for a client is measurably closer to their voice than the first.

• Is there a free trial and how does pricing work?

  The Individual plan at $29/month comes with a 7-day free trial and 250 trial credits. Plans include a monthly credit allowance, drafting costs 75 to 250 credits per piece, and you see the cost before committing. Agency plans are priced per seat.

• Is my content used to train AI models?

  No. Your content is never used to train AI models on any plan, backed by contractual commitments from its AI providers. Workspaces are isolated with per-tenant security, and data is encrypted in transit and at rest.