threads-humanizer

Threads Humanizer V3

Rewrites any Threads post or thread to remove the AI tells that human readers notice, and audits a finished draft against the 2026 Threads ranking checklist. Based on Wikipedia's "Signs of AI writing" taxonomy, the 2025-2026 stylometry literature, our own length-controlled Threads corpus (n=311), and Threads-specific patterns (the warm conversational register, the no-fold first line, the one-hashtag cap, repost-bait structure). V3 (2026-09): recalibrated on 2026 evidence. Vocabulary is scored by density, em dashes are capped instead of banned, forced rhythm is now a tell instead of a fix, and there is an over-correction guard.

What this skill does not do: it does not make text "pass" GPTZero, Pangram, Turnitin or Originality. Those are trained classifiers keyed on the instruction-tuning style signature; prompt-style "sound like a real person" rewrites are caught 92-95% of the time, and light mechanical rewriting raises detectability. On post-length text (under 300 words) detector scores are noise. The real value is elsewhere: expert human readers cite vocabulary (53%) and sentence structure (36%) as what gives AI text away, and Threads readers answer it with silence in a feed that ranks on replies. This skill removes what those readers react to.

What changed in V3

Evidence tier in brackets: [strong] = replicated across 2+ independent 2025-2026 studies or our own length-controlled corpus; [vendor] = single platform or vendor dataset; [weak] = one study or expert-panel report.

• Vocabulary moved from a delete-list to density scoring. The 2023-24 words (delve, tapestry, realm, journey) are decaying as humans avoid them [strong]. The durable 2026 markers are common words (significant, crucial, notably, comprehensive, insights, robust, leverage, foster, landscape, nuanced, streamline, elevate) plus grammar: nominalisations and "-ing" clause openers at 5.3x the human rate [strong]. In our Threads corpus AI vocabulary is nearly absent from top posts (1.3%) [strong], which is the point: it reads as a brand account in a feed built on people talking. One marker in a post is not a verdict. Three is.
• Em dash is no longer a tell; the density is. GPT-5.4 emits 1.43 per 1,000 words, below the 3.23 human baseline [strong]. On Threads specifically em dashes are rare in top posts (7%) and those posts earn 0.33x the median engagement [strong: corpus], so the cap here is tight: at most one per post, and none in a post that reads fine without one. Replace the excess with a comma, a colon, .., or a rewrite. Never a period (a split dash stacks fragments).
• Forced burstiness is the #1 2026 tell, not the fix. Mechanical long/short alternation is a learnable humanizer fingerprint [weak], and on Threads uniform rhythm wins at every length: sentence-length variance correlates negatively with engagement across the whole corpus (Spearman -0.31), with uniform rhythm ahead in the short, mid and long terciles [strong: corpus, length-controlled]. So Pass 2 never forces variance, on a single post or a thread. It only removes manufactured variance. "Short. Punchy. Done.", "No X. No Y. Just Z.", one-word posts for drama and "The result?" reveals are the current top tells.
• Rule of three is still a tell, at density. Tricolon runs at 2x expert-human rate across 2026 frontier models [strong], and on Threads it is rare in top posts (8%) and engagement-negative (0.28x) [strong: corpus]. Stacked or perfectly parallel triads and any second triad in a post get scrubbed. One natural triple with concrete items stays.
• Fingerprint injection was half wrong. Named entities and concreteness are supported [strong]; an odd-precision number with a referent in line 1 is the strongest opener. Bare numbers are not a discriminator, and inserted hedges and confessions backfire: performed hesitancy is 2x more common in LLM text, and sincerity announcements ("let me be honest", "unpopular opinion:" on a popular take) are a named 2026 tell [strong]. Pass 3 asks for a flat, dated, uncomfortable fact instead, stated in the warm register Threads rewards.
• Over-correction guard. Humanizer output has its own fingerprint [weak]. Pass 4 checks whether Passes 1-3 introduced the very patterns they were meant to remove. Edits are proportional to real problems. When in doubt, leave it.