- Home
- Skills
- Marketing & SEO
- Brand Voice Cloner
Works with the AI Agents you already use
Brand Voice Cloner
Turn 3 to 10 samples of a brand's real copy into a voice guide any AI agent can follow, in markdown plus a machine readable JSON constraint block, both rendered from one profile so they never disagree.
Free
Brand Voice Cloner
Example session with this skill installed
Build a brand voice guide from these samples of my copy, with tone scores, do use and never use lists, and before and after examples. Export the guide and the JSON constraints.
- Read your context and instructions
- Compiled the brand voice cloner
Brand Voice Guide
Tone scores (1 to 5)
Formality, Warmth, Boldness, Playfulness, Technicality, and Urgency, each with a justification and a verbatim quote from the samples.
Vocabulary
- Do use: phrases actually in the samples.
- Never use: corporate filler and tones absent from the samples, with reasons.
Rhythm and formatting
Sentence length, fragments, punctuation, formatting, contractions, repeated phrases, and filler terms present or absent.
Before and after examples
Five before and after pairs, using only vocabulary and rhythm the guide supports.
Exports
- voice-guide.md
- voice-constraints.json, rendered from the same profile, with a one-line instruction for pasting the JSON into a system prompt.
The profile is validated before export, including that every evidence quote appears verbatim in the samples.
Connects securely to your tools. The creator never sees your data.
About this skill
Turn 3 to 10 samples of a brand's real copy into a voice guide any AI agent can follow, in markdown plus a machine readable JSON constraint block, both rendered from one profile so they never disagree.
Give it samples of your own writing, labelled website, email, social, ad, or other, at least 150 words in total. Get tone scores with evidence quotes, do use and never use vocabulary grounded in your samples, rhythm and formatting notes, five before and after examples, and a voice guide plus JSON constraints you can paste into a system prompt.
Why it matters: AI drafts sound generic because the model has no grounded record of how you actually write. This skill builds that record from your real copy, with every claim tied to a verbatim quote, so future drafts can sound like you without inventing a voice.
What it does:
- Validates samples, with 3 to 10 samples and at least 150 words required, and reports unsupported files as skipped.
- Measures rhythm from the samples: sentence length, fragments, punctuation, formatting, contractions, repeated phrases, and filler terms present or absent.
- Scores six tone dimensions from 1 to 5, Formality, Warmth, Boldness, Playfulness, Technicality, and Urgency, each with a justification and a verbatim quote.
- Builds do use lists from phrases actually in your samples, and never use lists from corporate filler and tones absent from your samples, with reasons.
- Writes five before and after pairs, using only vocabulary and rhythm the guide supports.
- Validates the profile before export, including that every evidence quote appears verbatim in the samples, and that personal or sensitive details are an error.
- Exports voice-guide.md and voice-constraints.json from one profile, with a one line instruction for pasting the JSON into a system prompt.
Honest limits:
- It does not write the final marketing copy itself. Build the guide first, then write with it.
- It does not invent a voice for a brand with no existing copy. The guide must come from real samples.
- It does not score or lint new drafts against a guide in version 1. A drift detector and linter are planned for version 2. Offer a manual voice check instead.
- Version 1 stores no state between sessions. The exported files are the persistence.
- If you supply less than the minimum, it will not guess a voice. It will say what the minimum is, why it exists, and exactly what to add.
- Reading PDF samples uses the external pdftotext tool when installed. Without it, it asks for pasted text or a txt file instead.
- Sample pairs that diverge on two or more rhythm signals are flagged as possible second voices.
Requirements: An agent environment that can run Python 3 scripts with no Python dependencies for the measurable steps. PDF reading needs pdftotext if you use PDF samples.
Example prompt: Build a brand voice guide from these samples of my copy, with tone scores, do use and never use lists, and before and after examples. Export the guide and the JSON constraints.
What is in the package: SKILL.md, LICENSE, references including scoring rubric and profile schema, assets, and Python scripts for ingest, rhythm measurement, profile validation, and export, plus tests.
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
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
2 uses
Times used to date
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 23 days ago
- Free with an Agensi account
Requires