More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Own Voice Pro: Your Voice That Keeps Learning

    4

    Your brand voice for LinkedIn posts, cold email, Threads and freelance bids: voice profiles that keep learning from every correction.

    $8

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    5.0 (1 review)
    Own Voice Pro: Your Voice That Keeps Learning

    Own Voice Pro: Your Voice That Keeps Learning

    Example session with this skill installed

    Use my work voice, LinkedIn post: "I'm thrilled to share that six weeks ago our team moved away from weekly status meetings to an asynchronous shared document updated every Thursday! 🚀 Team members write two lines about what they shipped and what's blocked, and I review it in ten minutes instead of an hour-long meeting. We kept one thing: a short call when something is genuinely stuck. #Productivity #Leadership"

    • Read your context and instructions
    • Compiled the own voice pro

    Rewrote your draft with the

    work profile for this request only (your default stays personal). Critic: APPROVE on pass 2.

    Six weeks ago we dropped our weekly status meeting and moved to a shared doc everyone updates by Thursday.
    
    People write two lines about what they shipped and what's blocked. I read it in ten minutes instead of sitting through an hour.
    
    The one thing we kept is a short call when something is actually stuck.
    

    Profile · platform: work (inherits personal) · LinkedIn post · 308 / 3 000, counted with python3

    In your voice: dropped "thrilled to share" and the rocket (Never: excitement openers, emoji at work) · hashtags removed (C-004, no hashtags on LinkedIn) · "asynchronous shared document" → "shared doc", your Vocabulary · result kept in its own sentence (C-006).

    Before → after:

    • "Team members write two lines" → "People write two lines"
    • "instead of an hour-long meeting" → "instead of sitting through an hour"
    • "a short call when something is genuinely stuck" → "...actually stuck"

    Left alone: six weeks, Thursday, two lines, ten minutes, the one call you kept.

    The real test: does it sound like you? Change any line or paste the version you'll post. I'll fix the text, show a lesson card with the rule I'd learn, and save it to work only if you say yes. You have 9 unmerged corrections in work; at 10 I'll offer to consolidate them.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Own Voice Pro rewrites any text in the way you actually write.

    It starts with a ten-minute calibration built on a neutral baseline: you rephrase deliberately bland sentences your way, and since the baseline is known, every change you make is clean voice signal. No guessing from old samples that someone else already edited.

    Keep several voices in one skill: personal, work, a client brand. A work profile can inherit from your personal one and store only what differs. Say "use my work voice" for one request, or switch the default.

    Presets shape the text for where it's going: freelance bids and proposals, cold email, Threads, LinkedIn posts, connection notes and DMs, plus a template for any platform. Your profile decides the wording, the preset decides length and format, and hard limits are counted exactly.

    An independent critic checks every rewrite against your profile before you see it. In Claude Code it runs as a separate subagent that never sees the writer's reasoning.

    When you fix a line, the skill shows what it learned and saves it on your OK. The profile keeps growing in a dated corrections log and gets consolidated into clean rules.

    Honest scope: a voice tool. No "% human" scores, no detector promises. In testing, hand-written text scored 98% human on GPTZero and model rewrites 5–47%, so we don't sell that.

    Your profiles stay yours and stay out of the skill: in Claude Code they're saved to ~/.own-voice/ (or a folder you choose), on claude.ai they live in a Project. A profile from the free edition carries over.

    Works in English (Ukrainian is supported too). Plain Markdown, no scripts, no network calls. Free edition: github.com/volikablack/own-voice.

    Writing proposals? Freelance Bid Writer builds them only from facts you can prove, and this skill then puts them in your work voice.

    • Languages: English, plus Ukrainian.
    • Presets: Threads, LinkedIn (post, connection note, DM), email including cold email, freelance bids, plus a template for any other platform.
    • Setup: a ten-task calibration of about ten minutes; a child profile like work takes about five more.
    • Delivered as text through the Agensi agent connection: SKILL.md, references, two fictional example profiles, worked examples of the learning loop in EN and UK, and a critic agent file that is optional and only for local installs.
    • Your profiles stay outside the skill, in ~/.own-voice/ or a claude.ai Project.
    • Single-user license. Upgrading from the free Own Voice keeps your existing profile: Pro reads ~/.own-voice/profile.md, or you paste it.
    • No "% human" score and no AI-detector promises.
    • On claude.ai files don't save between chats: the skill hands you the profile text and you keep it in your Project.

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    5.0
    1 review
    5
    1
    4
    0
    3
    0
    2
    0
    1
    0

    Trust & safety

    Security scan

    Security scan passed 22 days ago

    Permissions

    Read Files
    Write Files

    File Scopes

    ~/.own-voice/**

    Reads the skill text your agent receives, your voice profiles and config.md in ~/.own-voice/ (or a folder you choose, Project knowledge, or text you paste). In Claude Code it writes ~/.own-voice/profiles/<name>.md, ~/.own-voice/config.md and, if you ask, ~/.own-voice/platforms/<name>.md; every write to a profile is shown to you first and saved only after you say yes. It never writes anything into a skill folder. On claude.ai it gives you the exact text to paste into your Project. If it finds a free Own Voice profile at ~/.own-voice/profile.md, it offers to convert it and leaves the original where it is. If an older local install keeps your profiles inside the skill's voice/ folder, it offers to move them to ~/.own-voice/ and does so only after you say yes. When a platform preset is used it pipes the text into a one-line python3 command to count characters, with no temporary file. It makes no network calls. When your agent can start subagents, it starts a critic subagent with the critic rules passed inline; the critic never sees the writer's reasoning.

    Listed1 month ago
    Updated22 days ago

    What's inside

    Frequently Asked Questions

    Corrections become saved rules through a lesson card, you can keep several profiles with inheritance, and there are platform presets with exact character counts. The free edition has one profile and no presets.

    No. It fixes the text first, then shows what it learned and asks. Nothing is saved without a yes, and silence counts as no.

    When a profile reaches 10 unmerged corrections, or when you ask, it merges them into clean rules, resolves contradictions and compresses the log. It shows you the diff before writing.

    The final text is counted with code, never estimated, and the critic rejects anything over a hard limit. If the platform's own counter disagrees, the counter wins.

    It makes no detector claims and prints no "% human" score. The only claim is that the text sounds like the person in your profile.

    Outside the skill, so updates never touch them. In Claude Code they're in ~/.own-voice/ (or a folder you name): one file per profile with its corrections log inside, plus config.md for the default. On claude.ai they live in a Project, and the skill hands you the exact text for each change. If an older local install keeps them inside the skill folder, the skill offers to move them.

    New in Social Media