A content team that argues with itself
Content Studio
Deep articles researched, rated, written, illustrated, and gated before you read them — then your whole published library graded overnight, with every score backed by a quote from the piece.
Content is where solo marketers drown: every asset needs strategy, drafting, editing, and brand consistency — usually across six tools. The Studio runs that pipeline in one place, grounded in the same ICP, personas, and voice the rest of the OS uses.
It closes the loop, too: a recurring sweep scores published content against your locked brand voice, customer-interview quotes flow into testimonials and drafts, and everything you upload becomes retrievable knowledge for future generations.
What makes it different is that every judgment is argued, not asserted. An idea gets a create / refine / skip verdict with a devil's advocate; a draft is checked against your own writing-habit dials and fixed by adding a person, not by fooling a detector; a chart is only drawn from a statistic the piece actually cites; a library grade must quote the sentence that earned it; and readers' comments become primary signal for the next round. Machines do the drafting, arguing, illustrating, and grading — overnight, on our own hardware, for a fraction of a cent per piece. You vote, lock, and publish.
What's in it
Deep articles, briefed for the reader
A brief that carries the reader's job-to-be-done, the depth (standard or deep), and an agent plan the AI pre-fills from your locked foundation — then Wow / How / Now writer modes, a critic that scores stance, evidence, voice, and information gain, and a citation verifier that checks every claim against its source before you see the draft.
Content ideas → briefs → drafts
A staged pipeline with AI drafting at each step and your judgment between steps.
Guided research
Steer a piece's research before a word is written: edit the questions, pin the domains it must use or never touch, attach your own sources — fetched, cited, and marked as yours. The finished piece shows exactly what grounded it.
Value rating
Every idea gets a create / refine / skip call across five axes before you spend a draft on it — with a built-in devil's-advocate paragraph that argues against its own verdict.
Your own material, with a receipt
Attach pasted text, links, or library documents to an idea — each marked as source, style, or reference. Links are fetched and extracted; the finished draft shows exactly which of your inputs it used.
Article visuals from the evidence
A per-piece visual plan: brand illustrations per section, charts rendered deterministically from statistics the piece cites (the model never draws numbers), a hero and section photos — every image stamped with IPTC provenance and one vote from regeneration.
Originality & human-voice gate
A plagiarism scan that gates publishing on real verbatim matches, an AI-probability score shown as advice (never a block), and a check of the draft against the writing-habit dials you set — then a human-voice pass that fixes AI tells by adding a person, shown as a diff you accept.
Content Review & Signal Ratio
Your whole published library scanned, fingerprinted, and graded overnight against your locked ICP, personas, and voice — five axes, each score backed by a verbatim quote from the piece, a keep / refresh / retire verdict, and one roll-up number with an honest denominator: the Signal Ratio.
Retrieval-ready by default
Answer-first leads, specific headings, a statistic and a quote per section, an FAQ block, a visible last-updated date, Article + FAQ schema, a Google preferred-source link — enforced by the writer, scored by the critic, and checked by a deterministic lint, so answer engines can quote you.
Coverage map & refresh trickle
Every graded piece mapped onto persona × funnel-stage cells — empty cells carry a one-click 'draft ideas for this cell' — and each week the review drafts refreshes for the few pieces it would move most. Drafts wait for your vote; nothing publishes itself, and your votes teach the next pass.
Onboarding burst
An empty studio fills itself: from your first signal or Compass run it ranks ideas, drafts the top three as deep dives, renders heroes from a starter brand kit, and lays out a first-fortnight calendar — labelled as grounded in a draft foundation and refreshed when you lock.
Reader comments → signal
Readers comment under your published posts (Turnstile, no tracking); a curator proposes approve / hide / spam with a devil's advocate; your approval publishes the comment and mints it as primary-tier signal that feeds the next round of ideas.
Content Readiness Scorecard
The eight fundamentals that still decide whether you get found — positioning, ICP, intent, a sound site, topics, original content, authority, thought leadership — graded from data the OS already holds, each with a quote as evidence, honestly 'not scored' where a source is missing, confirmed pillar by pillar by you, trended after every re-lock.
Content agents you can steer
Ideator, drafter, illustrator, grader, curator, visibility tracker — every step runs as a named agent you can pause, pace, and cap, with its spend in view. Drafts and proposals never publish themselves.
Book Studio
When your foundation locks, write the book: archetype proposals grounded in your signal, chapter-by-chapter drafting in your voice with a continuity editor, and a KDP-ready export — plus a done-for-you Amazon publication package (paperback, Kindle, audiobook).
Brand-voice adherence sweeps
Scheduled checks that score published content against your locked voice.
Campaigns
Multi-asset campaign planning tied to playbook plays.
Upload Center & knowledge base
Decks, docs, and transcripts parsed into evidence the AI can cite.
AI & human transcription
Fast AI transcription by default; certified human transcription packs when accuracy is contractual.
Signal
A quality-tiered evidence stream — what the AI knows about your market, auditable and enrichable.
Works with
Put Content Studio to work on a real client.
$0 base fee through the open beta. Upgrade when the engagement lands.