Brand Style & Terminology Linter
Paste a draft, check it against your own style guide, and get every banned term, wrong product name, and spelling slip flagged with a suggested fix - approved by an editor before anything changes.
An internal web tool where your team pastes a draft, sees every style-guide deviation flagged with a suggested fix, accepts or dismisses each one, and walks away with corrected text plus a logged report - rules editable as a simple sheet.
Before you start
- A free Vercel account
- A free Supabase account
- A free Resend account (for optional report emails)
- Your style/terminology rules in a spreadsheet or in your head (term, preferred form, banned form, note)
The problem this kills
"E-commerce" or "ecommerce"? Is it the Pro plan or the Premium plan this quarter? Did someone call it a "synergy" again? When ten people write copy, your style guide lives in a forgotten PDF and the rules slip through anyway. You catch the misses by re-reading every draft by hand - or you don't catch them, and a wrong product name ships to thousands of inboxes.
A style guide nobody enforces is just a document. This tool turns your guide into a check that runs on every draft, every time, in seconds - so "off-brand" stops being something you hope people remember.
What you'll build
An internal web app, just for your team, that does one job extremely well:
- Your editor pastes a draft (or its title and version).
- The tool runs it against your rules sheet - approved product and feature names, banned jargon, preferred spellings, capitalization rules.
- It lists every deviation it finds, each with the exact spot, the rule it broke, and a suggested fix.
- The editor accepts or dismisses each flag - nothing is ever silently rewritten.
- You get clean corrected text plus a logged report of what was found and what the editor decided.
- Spotted a rule you don't have yet? Add it from the flagged miss in one click, and it applies to every future draft.
It supports whole-word matching (so "cat" doesn't flag "category") and case-sensitive rules (so "iOS" stays "iOS"), and it imports and exports your rules as a plain spreadsheet.
What's inside the Implementation Plan
The plan is a runbook you paste into an AI coding agent (Claude Code), which then builds the tool with you step by step - no coding required from you.
It opens by interviewing you about your business. Before a single line is built, the plan has the agent ask about your current review process, where your style rules live today, how your product and feature names are spelled, your real banned terms and preferred forms, and your messiest edge cases. Then it reads back a short tailored spec and waits for your thumbs-up. The result is a linter tuned to your brand voice - not a generic word list.
From there it walks through: setting up the database and login, importing your rules sheet, the paste-and-check screen, the accept/dismiss review gate, the "add a rule from a miss" flow, the corrected-text export, and the logged report - each step ending in a ready-to-paste prompt.
The governance it includes (this is the point)
This isn't a toy. Every build includes the controls that make an internal tool safe to actually use:
- Login so only your team can open it.
- Row-level security so each organization only ever sees its own rules and reports.
- A complete audit trail - who checked which draft, what was flagged, and what they accepted or dismissed, with timestamps.
- A human-in-the-loop approval gate - the AI drafts suggestions; a person reviews and approves each one; only then is the corrected text produced. Nothing is auto-rewritten.
- Duplicate guards - the same draft title and version can't be logged twice by accident.
Who it's for
Content managers, brand managers, and marketing ops people who are tired of "ecommerce vs e-commerce," wrong product names, and stale jargon slipping through into published work - and who want their style guide to actually get enforced without policing every draft by hand.
You've got this. Paste the first prompt and let the agent interview you.