- Home
- Skills
- Writing & Content
- Own Voice: Rewrite Text in Your Own Voice
Works with the AI Agents you already use
Own Voice: Rewrite Text in Your Own Voice
Humanize AI drafts into your own writing style: a five-minute calibration, then rewrites that sound like you, in English or Ukrainian.
Free
Own Voice: Rewrite Text in Your Own Voice
Example session with this skill installed
Rewrite this in my voice, the AI draft sounds nothing like me: "Hi Mark, I wanted to let you know that the website will take a little longer than originally planned. The payment form had an issue on mobile that required careful attention. I apologize for any inconvenience. The new delivery date is Friday the 14th, and I'll share a preview link before then."
- Read your context and instructions
- Compiled the own voice
Rewritten with your saved voice profile. Critic: APPROVE on pass 1.
Hi Mark,
the site will be a few days late, the payment form broke on mobile and I want it working properly before you see it. New date is Friday the 14th, I'll send you a preview link before then.
In your voice: lowercase start after the greeting, the way you write · reason and fix in one flowing sentence joined with commas, since you rarely use full stops mid-thought · dropped "I apologize for any inconvenience" (your profile: you don't apologize for a delay you're already fixing) · no dashes.
Before → after:
- "required careful attention" → "I want it working properly before you see it"
- "The new delivery date is Friday the 14th" → "New date is Friday the 14th"
- "I'll share a preview link" → "I'll send you a preview link"
Left alone: the client's name, the mobile payment form issue, Friday the 14th, the preview link.
The real test: read it. Does it sound like you? If one line doesn't, tell me which one and I'll fix it and add the correction to your profile. No detector score here; this is a voice tool.
Connects securely to your tools. The creator never sees your data.
About this skill
Own Voice rewrites text in the way you actually write. Hand it an AI draft that doesn't sound like you, a rushed email or a stiff paragraph, and it gives it back with your rhythm, your words and your punctuation.
It starts with a five-minute calibration: ten short tasks where you rephrase deliberately bland sentences your way. Since the bland version is known, every change you make is clean voice signal, and the skill saves it as your voice profile.
An independent critic checks each rewrite against that profile and sends back anything that sounds like an assistant instead of you.
Honest scope: this is a voice tool. It doesn't print "% human" scores and doesn't promise to beat AI detectors. We tested that: hand-written text scored 98% human on GPTZero, model rewrites 5–47%, so we don't sell it.
Works in English (Ukrainian is supported too). Your profile is kept outside the skill: ~/.own-voice/ in Claude Code, a Project on claude.ai. Plain Markdown, no scripts, no network calls. MIT, source on GitHub: github.com/volikablack/own-voice. Need several voices, platform presets (LinkedIn, Threads, email, bids) and a profile that learns from your corrections? See Own Voice Pro.
- Languages: English, plus Ukrainian. Any other language, it says so and stops.
- Setup: a calibration of ten short tasks in three groups, about five minutes, and you can stop after any group.
- Delivered as text through the Agensi agent connection: SKILL.md, the writer, critic and calibration references, pattern lists for EN and UK, an example profile, and a critic agent file that is optional and only for local installs.
- Your profile stays outside the skill, in ~/.own-voice/ or a claude.ai Project.
- MIT license. Source on GitHub.
- It never prints a "% human" score and makes no AI-detector promises.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
This skill has been used 7 times.
Trust & safety
Security scan
Security scan passed 22 days ago
- Free with an Agensi account
Permissions
File Scopes
Reads the skill text your agent receives and your voice profile (~/.own-voice/profile.md, Project knowledge, or text you paste). Writes one file, ~/.own-voice/profile.md (or the folder you choose), after the calibration and after corrections you confirm, and always shows you the profile before saving. It never writes anything into a skill folder. If an older local install keeps voice/profile.md inside the skill folder, it offers to move it to ~/.own-voice/ and does so only after you say yes. When your agent can start subagents, it starts a critic subagent and passes it the critic rules inline. It runs no scripts and makes no network calls.