Paste 3-5 of your existing posts. AI extracts your voice DNA — tone traits, sentence patterns, vocabulary signatures.
Paste 2-5 pieces of your existing writing (LinkedIn posts, tweets, newsletters, blog excerpts — anything you wrote yourself). AI extracts the voice DNA underneath — patterns you use without noticing.
Persona Brief Builder takes your stated preferences and structures them. Voice Analyzer reverse-engineers your actual voice from your existing writing — including patterns you use without noticing. Most creators are surprised by what the analyzer surfaces because the stated voice and actual voice diverge.
One sample is anecdote, not pattern. Patterns emerge across multiple pieces. 3+ samples lets the analyzer distinguish a one-off phrasing from a signature move. More samples = more accurate, up to 5 — beyond that, you're mostly confirming what you already know.
Pieces YOU wrote (not collaborator drafts). Longer-form is better — LinkedIn posts, blog drafts, email newsletters, Twitter threads. Avoid short tweets in isolation; they're too constrained to reveal voice. 200-1500 words per sample is the sweet spot.
Paste it into your Persona Brief. The "rhetorical moves" and "vocabulary fingerprint" sections become AI generation rules — Kompozy reads them at every generation and matches your voice. The "anti-patterns" become hard rules in your prompts. Most creators see 40-60% better voice fidelity after this pass.
Kompozy automates everything this tool estimates — multi-format generation, brand voice, scheduling across 9 platforms. Free 500-credit trial, no card required.