Tastemark - Brand Voice Scoring for AI Content
Scores AI-generated marketing copy against your brand guidelines and rewrites what doesn't match, so every piece sounds like you wrote it.
AI writing tools produce generic, off-brand content that marketing teams must manually revise through costly editing cycles. No existing tool acts as a dedicated quality-control layer that scores and fixes brand voice issues across content from any source.
Build This Idea
Run Claude Code then copy and paste this command
Sign In to Build
Join with a free account
Claude Code will scaffold the full project for you based on the idea spec, tech stack, and features
Talk to Claude Code to edit features, add integrations, or customize anything in your new project
The Business
$8-12B
Market Size
$39-$149
Global
Worldwide Potential
Customer
Marketing teams and content managers at mid-market B2B/D2C companies (50-1,000 employees) and agencies managing multiple client voices who already use AI writing tools.
Pricing
Tiered SaaS subscriptions: Starter at $49/mo (1 brand, 50 scores/mo), Pro at $99/mo (3 brands, 300 scores/mo), and Agency at $249/mo (unlimited brands, 1,000 scores/mo, API access). Usage-based overages at $0.10 per additional score. Annual plans at 20% discount to improve retention.
$9.4M
Estimated Annual Revenue
20,000 customers at $39-149/mo
10% market capture
Features
Upload style guides, paste sample copy, and define tone attributes to train a per-brand voice model with adjustable dimensions like formality, humor, and jargon level.
Paste or import AI-generated copy and receive a 0-100 brand alignment score with line-by-line annotations showing which phrases deviate from guidelines and why.
One-click rewrite of flagged sections to match brand voice while preserving meaning, with before/after diff view for human approval.
Agencies and multi-brand companies can maintain separate voice profiles and switch between them. Each workspace has its own scoring model.
Chrome extension that overlays scoring and rewrite directly inside Google Docs, Notion, Jasper, and Copy.ai for in-workflow brand checking.
REST API endpoint that accepts text and returns scores plus rewrites, enabling integration into automated publishing workflows.
Track average brand scores over time by team member, content type, and source tool to identify where voice drift happens most.
Define required terms, preferred phrasings, and banned words that get flagged automatically alongside voice scoring.
Sign In to View
Join with a free account
Tech Stack
apis
OpenAI GPT-5.3
Primary LLM for voice scoring analysis, rewrite generation, and brand guideline interpretation
Anthropic Claude
Secondary LLM for nuanced tone analysis and as a fallback for complex rewrite tasks
OpenAI Embeddings
Generate text embeddings from brand samples to build the voice similarity model in Pinecone
backend
Next.js API Routes
Handle scoring requests, brand profile CRUD, and orchestrate LLM calls
Supabase Edge Functions
Lightweight serverless functions for webhook integrations and async rewrite processing
hosting
Vercel
Deploy Next.js app with edge functions, fast global CDN, and seamless preview deployments
database
Supabase
PostgreSQL for brand profiles, scoring history, user management, and team analytics with built-in auth
Pinecone
Vector database to store brand voice embeddings from sample copy for semantic similarity scoring
frontend
5 Day Sprint UI
Component library built on shadcn/ui and Tailwind for rapid, consistent UI development
Next.js
React framework for the dashboard, onboarding flows, and diff/rewrite interfaces with SSR
Sign In to View
Join with a free account
Frequently Asked Questions
Start with your own idea and setup an AI business today