Pre-send scan every paid issue, free post, Note, and About page before you hit publish in Substack, Beehiiv, Ghost, or ConvertKit. Sentence-level highlights show which lines drift into AI-template prose so you can rewrite in your voice rather than guess. Built for solo newsletter writers protecting voice consistency across 50-plus posts and for multi-author publications running a shared editorial bar. Free to try. No card.
Solo Substack writers running a single paid publication, multi-author masthead teams editing across guest contributors, and serial essayists shipping into Beehiiv or Ghost all share the same need: a fast pre-send scan that protects voice consistency before an issue lands in a subscriber inbox.
Newsletter writing is one of the few formats where voice is the product itself rather than a side effect. Affiliate blogs sell placements, SaaS landing pages sell features, and an SEO blog sells search rank. A paid Substack sells the way you write. Pre-send scanning fits the format because the asset being protected sits in the sentences, not in the topic.
One weekly issue, occasional Notes, the rare bonus essay. Pro at $14.99 a month yearly gives unlimited scans, 10,000 character pastes covering a typical 1,800-word issue in one pass, and 90-day history covering roughly a quarter of issues. The integrated AI rewriter handles the stubborn paragraph in the middle of an essay that flags every time without forcing a voice rewrite of the whole piece.
Three to ten writers shipping under one masthead, with a managing editor enforcing a minimum quality bar before each issue goes out. Business at $29.99 a month yearly unlocks five seats with shared scan history, REST API access for workflow automation, an audit log so the editor can see who scanned what when, and white-label PDFs branded to the publication for guest-contributor records.
Writers building a recognisable persona across 50-plus posts in the same column. Voice consistency is the asset, and a stable Authenticity Score across the archive is the diagnostic. Use the scan history to spot which issues drifted, which series carried voice forward cleanly, and which formats you ship strongest in.
Paid posts, free posts, Notes, the About page, and recommendation write-ups carry different stakes and score differently. Read the score in context of the surface rather than chasing a single number across every type of work you publish.
The highest stakes on any publication because every paid post carries the subscription line item on a reader's card. Self-scan every paid issue before hitting Send. Target an Authenticity Score of 80 or higher on long essays and rewrite anything below 75 in the sentence-highlights list. The intro and the closer are the usual suspects on essays because both default to stock phrasing when the draft moves fast.
Free posts must read like the paid product because free readers are deciding whether your voice is worth paying for. A free post that smells AI burns a future subscription before it ever converts. Scan with the same threshold you use on paid issues. Most working writers find that the free top-of-funnel post is the highest-leverage place to invest a few extra minutes of voice polishing.
Substack Notes, Threads-style short-form, and quote-stack posts sit on public social feeds where AI flavour stands out faster than in a long essay. A run of templated Notes taints how readers receive your next paid issue. Paste each Note into TextSight before posting; most Notes fit inside a single scan window with room to spare.
The About page is the highest-leverage page in your archive for converting free readers to paid. Scan it once a quarter and after any rewrite. Recommendation write-ups appear on other publications and act as ambassadors for your voice in front of readers who have not signed up yet; a templated blurb costs you cross-publication conversions.
The automated welcome and renewal emails are the first and last contact each subscriber has with your voice. If those read AI, the framing of every issue afterwards is coloured. Scan once, rewrite once, and they carry your voice forward without further attention for a year.
Pro at $19.99 a month standard, $14.99 a month on yearly, is the right fit for solo Substack writers and serial essayists. Business at $39.99 a month standard, $29.99 a month on yearly, fits multi-author publications running a shared editorial bar. Full details on the pricing page.
Billed $89.88/year — Save $30
Billed $179.88/year — Save $60
Billed $359.88/year — Save $120
Yearly billing saves 25%. View full pricing →
Most newsletter writers draft outside the publishing tool, then paste the final version into Substack, Beehiiv, Ghost, or ConvertKit at send time. The scan step slots in cleanly between drafting and pasting, and the whole loop adds 30 to 90 seconds to a clean issue.
Notion, iA Writer, Apple Notes, Google Docs, Obsidian, Bear, or whatever your drafting tool is. Keep that flow. AI tools are fine upstream during outlining or when you are stuck on a section; the scan step enters only when the prose is final and you are about to load it into the editor.
The app returns an Authenticity Score from 0 to 100 plus a list of sentences flagged as AI-template. The score gives you a personal publish floor (most working writers settle on 80). The sentence list tells you exactly which lines to rewrite, so you do not waste a voice-pass on prose that already reads like you.
The point is not to write differently. The point is to write in your existing voice in the places the draft slipped. If a sentence flagged because the rhythm went template, rewrite it with your normal cadence. If a sentence flagged because the transition felt formulaic, rewrite the transition the way you would have in an early issue from your archive.
Run the cleaned draft through TextSight a second time. If the score is at or above your floor and no critical sentences remain highlighted, paste into the Substack, Beehiiv, Ghost, or ConvertKit editor and hit Send. The integrated AI rewriter on Pro handles the stubborn paragraph in the middle of a piece without forcing a full voice rewrite.
Engaged readers learn your voice fast. A paid subscriber reads everything you publish, often more than once. When a paragraph reads AI, they catch it before the first scroll. They might not comment. They will quietly resent the issue. Multiply that across a paid list and you have a churn engine running in the background.
A WordPress affiliate blog can sell placements or ads; a paid newsletter sells your voice directly. The subscription line item on a reader's credit card is paying for the way you write. If three issues in a row read AI-template, those readers might not cancel immediately, but they stop opening, they stop sharing, and they let the renewal lapse. The decay is silent because no platform surfaces a churn-on-AI-suspicion signal, but it shows up in your renewals after a quarter.
At 200 paid subscribers at $5 a month, one subscriber who stays because an issue sounded like you covers Pro for an entire year. The break-even is one retention save. At 1,000 paid subscribers the break-even is roughly one save per quarter; at 5,000 paid subscribers the break-even is roughly one save per year.
Publishing platforms across the board have rolled out AI labeling and signal features over the last two years. The mechanics keep shifting, but the trend is consistent: posts that read AI-template at the classifier level can be demoted in recommendations, in cross-publication surfaces, and in feeds. A clean pre-send scan generally correlates with not getting flagged by the platform afterwards.
The newsletter network is built on restacks and cross-publication recommendations. Distinct voice is what gets restacked. Template prose disappears in the feed. Pre-send scanning is partly a hedge against platform labeling and partly a positive bet on the voice signal that the network amplifies.
Long-running newsletter series live or die on voice consistency. A reader who has been on the list for two years knows your sentence rhythm, your favourite hedges, your specific way of opening a paragraph. The scan history is the diagnostic for whether you are holding voice across the archive.
Weekly publishers run the scan as the last step before Substack, the same way a daily journalist runs spellcheck. Bi-weekly publishers tend to scan two or three times across a draft cycle, once after the outline-to-prose pass and once before send. Monthly essayists treat the scan as a deep-work step the day before publish, with time built in for a voice-pass on flagged sentences.
For columns running 50-plus posts under the same name, the scan history is the long-term diagnostic. A stable Authenticity Score across the archive means voice held. A score that drifts down over six months means the series picked up a templated cadence somewhere, and the highlights will show where. Pro's 90-day history covers roughly a quarter of weekly issues, which is the right granularity for spotting drift early.
Over 10 to 15 issues you internalise which phrasings your voice does not use. The first month feels heavy because every flagged sentence is a discovery. By month three the scan becomes a 30-second reflex and the voice-pass becomes muscle memory. The skill you build is not "writing to pass detection"; it is recognising drift in your own prose, which is the same skill that long-running newsletter writers build by hand over years.
Multi-author publications hand each guest contributor the publication's minimum Authenticity Score (typically 80) as part of the brief. Contributors scan their drafts before submission and include the score in the email. The managing editor scans on receipt and rejects below-floor drafts back to the contributor with the sentence highlights attached. The audit log on the Business tier keeps the records straight quarter to quarter.
A single percentage is not a fix path. The TextSight result panel shows which sentences reacted and why, with paragraph-level rollups for longer essays, so you can edit specific lines instead of rewriting the whole issue.
Every sentence is colour-coded by its own AI-likeness score. Red sentences clustered in one paragraph are a stronger signal than scattered yellows. Scattered yellows in otherwise structured prose often just mean a stock transitional phrase. You read the pattern, not just the headline number.
Longer essays get paragraph-level rollups so you can see which paragraph is dragging the headline score. The intro and the closer are the usual suspects on newsletter essays because both default to stock phrasing under deadline pressure. Targeting the lowest paragraph first is the fastest way to lift the issue.
Perplexity is how predictable your word choices are to a language model. Low perplexity reads AI-like. The signal is shown per-sentence on Pro, which is the diagnostic context you need to decide whether a flag is real AI residue or a particularly well-rehearsed turn of phrase you have built into your voice deliberately.
Burstiness is how much your sentence length and structure vary across the piece. ChatGPT defaults to uniform medium-length sentences. Real newsletter writing has bursty rhythm: one short sentence, one long, one fragment. Low burstiness across an entire issue is the classic AI fingerprint and the one engaged readers learn to spot first.
CMS-agnostic blog workflow, sentence highlights, and the pre-publish loop.
For bloggers →The broader writer-audience page covering client deliverables and brand voice.
For writers →Light, Balanced, and Maximum modes for fixing flagged passages without losing voice.
Read the guide →Free, Starter, Pro, Business. Yearly billing saves 25%. Solo to publication tiers.
See pricing →Free to try. No card. Pro at $14.99 a month on yearly for solo writers; Business at $29.99 a month on yearly for multi-author publications.