- Holder
- English voice
- Language
- English
- Reference
- Sealed exam · 45 people
- Measured on
- —
Stop telling AI to “sound like me”.
Your voice, wherever AI writes.
Build your voice once. Take it into your AI tools. See the moment a text stops sounding like you.
Create my voice for free- Free to start: 10 credits
- No card required
- Your voice stays yours
stamp
VoiceEnglish voice · EN
Texts53 of 127
0/221
generic AI texts taken for the author · 11.10.2026
The passport gathers the texts that give your voice its material.
More content.
Less of you.
Five generated drafts. Five openings everyone has read before.Generic writing didn’t wait for AI. AI made it free, and the same for everyone.
In today’s fast-paced world… AI is transforming every industry, and it’s crucial…
Here are 5 ways to… stay relevant this year.
It’s important to note… that your content strategy needs a rethink.
Unlock the power of… AI to boost your productivity.
Let’s dive into… this week’s key takeaways.
Sounding human isn’t sounding like you.
So we tried hiding the AI:
AI detectorsHumanizers“Make it sound more human”
Still, nobody recognizes you.
Not an AI detector · not a humanizer
A prompt is not a voice.
A signature is recognized by the hand that made it, not by its shape. A prompt describes the shape; our engines learn the hand.
Sound like me: warm, direct, a little funny, no jargon.
- Your style rules
- Your cadence
- Your vocabulary
- Your convictions
- Your sub-voices
- Your learned rejections
What Kaneme is not
- Not a way to slip past AI detectors.
- Not a way to hide AI use — we show where the words came from ourselves.
- Not a verdict on writing with AI: that’s the use case, not a fault.
- Not a way to watch someone who hasn’t agreed to it: a measurement needs that voice’s own texts.
You don’t write an email the way you write a post.
One voice, several sub-voices. Kaneme spots the formats where your writing changes, and suggests one for each.
Hi Claire,
Thursday at 10, writing workshop. We start from your drafts, not from theory. You leave with a finished piece.
Coffee’s on me.
Shaku
Illustrated example, not a real account.
A sub-voice starts switched off. It is ready only after 3 human texts and 600 words in the format, and you are the one who switches it on.
Your voice can be checked.
Not guessed.
Your assistant writes with your voice. The text comes back measured against your reference.
Write the LinkedIn post announcing Thursday’s workshop.
In today’s ever-evolving landscape, it’s crucial to stay ahead. That’s why we’re thrilled to announce our game-changing workshop this Thursday. Let’s dive into the challenges of tomorrow!
prepare_voice_contextYour voice · version 2Your assistant receives your voice, its version and its limits.Thursday. Workshop. Your drafts, not theory. You walk out with a finished piece. Coffee’s on me.
- Short sentences
- “Game-changing” dropped in version 2
- A concrete detail to close
verify_authorshipVerdictA text comes back measured against your reference, with its probability.Release reading
Nemi EN 1 · sealed exam · protocol 1.0
- Engine
- Nemi EN 1
- Read on
- 11.10.2026
- People never seen
- 45 · 3 genres
Generic AI texts · 221
The authors’ own texts · 127
- Blogs & newsletters
- 21 / 70
- Literature
- 30 / 53
- Academic
- 2 / 4
Below the bar, Kaneme says less rather than guess wrong.
Every figure on this slip comes from one frozen read of authors held out of training.
Paper outside, chip inside
One voice.
Every AI.
Models are interchangeable. Your voice isn’t.
- 1The paperWhat you read: your voice, its rules.
- 2The antennaYour AI reads your voice, with your permission.
- 3The chipOur engines learn and measure your voice.
If your tool speaks MCP, it speaks Kaneme.
- Claude
- ChatGPT
- Gemini CLI
- Copilot Studio
- Grok
- Perplexity
- Cursor
- Claude Code
- Codex
- n8n
- + any MCP client
And for everything else, your own code, through the API: pip install kaneme
- Word
- Google Docs
- Notion
Writing in Word, Google Docs or Notion? Your AI tools already reach them. Your voice travels with them.
https://kaneme.com/api/mcpPaste it into your tool. It asks you to sign in, that’s all: nothing to generate.
The API and MCP: one account, and your credits.
Free, Credits, Business.
Your voice stays yours.
10 credits and 10 free portages every month, no card required. Purchased credits never expire. The monthly allowance resets without rollover.
- 1 voice included.
- Measurement, editor, REST API and MCP use the same prices.
- 2 voices included, then 5 credits per voice.
- Optional automatic top-up, with your threshold and monthly cap.
- 2 seats included, then $5 per month per seat, returned as credits on the shared balance.
- A budget per seat, personal or brand voices.
- Invoicing for larger accounts.
Kaneme writes in your tool: no quota, 3 credits per text, for every format and channel.
Your voice should stay yours.
Create your voice: 100 words are enough, no card needed.
Questions people ask us
Is Kaneme an AI detector?
No. An AI detector asks whether a text came from a model without knowing its author. Kaneme asks a different question, and only after you provide your own writing: does this text carry your voice? AI assisted writing is a supported use case; Kaneme measures whether your identity remains recognizable.
Is Kaneme a humanizer?
No. A humanizer optimizes text to evade a detector and conceal its origin. Kaneme does not test against AI detectors or offer an undetectability setting. When Kaneme generated the wording, that origin appears beside the identity result. A text can carry your voice and have been drafted by a model; both facts remain visible.
Can I try Kaneme for free?
Yes, in two ways, with no card. The “Test my voice” page returns a stamp without an account from three of your texts, without keeping them. Your passport starts with 100 words, and a free account then saves your voice. 10 credits and 10 free portages each month on Free and Credits accounts. Reset on your signup anniversary day, without rollover; the allowance is spent before purchased credits.
Can I connect Kaneme to my tools and agents?
Yes. The same identity and measurement operations are available through the REST API and MCP server. An agent can read a voice, measure a text, realign it and return a verdict without inventing its own definition of your style. API and MCP are open to every Kaneme account, Discovery included, and use the same credit balance as the app.
Are my texts used to train a model?
No. Your texts build your own voice reference. They are not added to a shared training corpus, and the verdict engine is never recalibrated on AI generated drafts. That separation protects the meaning of later measurements.
How is this different from ChatGPT?
ChatGPT drafts from an instruction. Kaneme first builds a reference from your own writing, then measures whether a draft still carries that voice. The same identity works in ChatGPT, Claude and your agents, so you can change models without rebuilding your voice or losing the independent verdict.