StyleMCP

StyleMCP provides an executable workflow for validating and rewriting AI-generated text to match a brand’s voice. You can access it as a REST API, integrate it as an MCP server for agents like Claude, run checks locally with the CLI, or wire it into GitHub Actions to catch off-brand copy during pull requests. It helps you enforce brand language, reduce risky or inconsistent copy, and automate copy governance across your stack.

How to use

You connect to the StyleMCP server using either the HTTP API or an MCP client integration. Use the API to validate or rewrite text from any application, or run the server locally via the MCP client for direct, in-context branding checks. When you need ongoing governance in your development workflow, enable automatic validation in your CI workflows or integrate the server into your existing agent ecosystem to ensure every generated message aligns with your brand.

How to install

Prerequisites: you should have Node.js and npm installed on your machine. You may also use Docker if you prefer containerized deployment.

Self-hosting with Docker:

# Clone the repository or pull the latest image
# Set up environment with a generated API key
echo "STYLEMCP_API_KEY=$(openssl rand -hex 32)" > .env

# Run with Docker
docker compose up -d

# Check health
curl http://localhost:3000/health

Manual self-hosting (non-Docker):

# Install dependencies
npm install

# Build the server
npm run build

# Start the server
npm start

If you plan to use the REST API directly from the internet, you can validate with the example endpoint shown in the quick start, or host your own instance and point your clients to that URL.

Additional configuration and usage notes

Two common ways to run StyleMCP are through an external API endpoint and through an MCP client integration. The API lets you validate and rewrite text from any application, while the MCP client option lets you embed brand rules directly into your AI agent workflows.

API and packs overview

StyleMCP exposes endpoints for validating text, rewriting text to match brand voice, and retrieving pack details. Packs are YAML files that define your brand rules, including vocabulary, forbidden words, patterns to avoid, and CTA guidelines. You can also query available packs and voice guidelines to tailor the checks to your needs.

Security and access

Protect your API key and restrict access to trusted clients. When running in production, consider using HTTPS with proper authentication and rate limiting to prevent misuse.

Troubleshooting and tips

If validation results show unexpected violations, verify that the text is evaluated against the correct pack. Ensure your API key is valid and that your client is pointing to the correct API URL. For local runs, confirm that the server process is running and listening on the expected port.

Advanced usage and examples

The style packs include vocabulary rules, forbidden words, patterns to avoid, and CTA guidelines. You can import your own pack, adjust tone and vocabulary, and use copy patterns to maintain consistency across campaigns.

Available tools

validate_text

Validate input text against your brand rules and return a validity status, score, and any rule violations.

rewrite_to_style

Rewrite given text to conform to the defined brand voice and tone rules.

get_voice_rules

Retrieve voice and tone guidelines for current brand style packs.

get_copy_patterns

Fetch approved copy templates and reusable patterns from the active pack.

get_cta_rules

Obtain CTA guidelines to ensure compliant action prompts.

get_tokens

Return design tokens and related assets used by branding rules.

list_packs

List all available style packs and their metadata.