2. Description of Service

Design.md Generator extracts design-system information from three sources:

• Public websites — via headless browser (Playwright on Node) or edge-based fetch + CSS parsing (Cloudflare Workers)
• GitHub repositories — via public GitHub API, parsing config files and stylesheets from public repos
• Visual screenshots — via client-side palette extraction, optionally enhanced by AI vision
• Text briefs — via natural-language prompts to external AI providers (when enabled)

The Service generates DESIGN.md documentation files intended for use with AI coding agents. Processing can be fully algorithmic (no AI) or AI-assisted depending on instance configuration.

3. Authentication

The Service offers optional GitHub OAuth authentication. By signing in:

• You authorize us to access your public GitHub profile (ID, login, avatar) and list your repositories.
• You acknowledge that your GitHub OAuth access token is stored in our database encrypted with AES-GCM-256 to enable repository listing and session persistence.
• You are responsible for maintaining the confidentiality of your session. Logout is available at any time.

4. Permitted Use

You agree to use the Service only for lawful purposes. Specifically:

• You may only analyze websites and repositories you have the right to access.
• You may not use the Service to scrape private, authenticated, or restricted content without authorization.
• You may not attempt to bypass rate limits, SSRF protections, or bot detection.
• You may not use the Service to probe internal networks or infrastructure.
• You may not use extracted design systems in ways that violate the target site's Terms of Service or intellectual property rights.

5. Rate Limiting & Fair Use

To ensure service availability, we enforce per-IP rate limiting. Limits differ by endpoint:

• Standard extraction (URL / GitHub / stream-extract / generate-design-md): hourly limit per IP, configurable by the operator (default: 10 per hour).
• AI generation (analyze-image / generate-design-md-ai / generate-brief): stricter limit (half of standard quota) due to API costs.
• Global API limit: twice the hourly rate per 10-minute window.

Excessive or automated abuse may result in temporary blocks. Cached results (KV) do not count against your quota.

6. Algorithmic vs AI Processing

The Service operates in two modes depending on instance configuration:

6.1 Algorithmic Mode (Default / AI Disabled)

When AI is disabled, all processing is algorithmic: CSS parsing, Tailwind/Bootstrap class inference, color bucketing, and token normalization. No data is sent to external AI APIs.

6.2 AI-Assisted Mode (Optional)

When enabled, extracted tokens, screenshots, and/or text briefs may be sent to external AI providers for markdown enhancement or visual analysis. AI features include automatic fallback chains (primary model → fallback models → text-only fallback) to maximize availability.

7. Intellectual Property

7.1 Generated Content

DESIGN.md files generated by the Service are derived from publicly visible design tokens of the analyzed source. We do not claim ownership over third-party visual identities, brand assets, or design systems.

You are free to use generated DESIGN.md files for personal, educational, or commercial projects. You remain responsible for ensuring compliance with the intellectual property rights of the original source.

7.2 Trademarks

All trademarks, logos, and brand names mentioned or analyzed belong to their respective owners. Their inclusion in generated documentation does not imply endorsement or affiliation.

8. Known Designs & Third-Party Content

The Service references pre-generated DESIGN.md files from community catalogs. We acknowledge and rely on the following open-source collections:

Additional community mirrors exist (e.g., Japanese UI collections). Discover more at github.com/topics/design-md.

9. Data Security & Risk Disclosure

We follow industry best practices: HTTPS/TLS for all communications, SSRF protection, bot detection, encrypted session cookies, and strict Content-Security-Policy headers.

However, you acknowledge and accept that while your GitHub OAuth access token is stored in our database encrypted with AES-GCM-256 to enable repository listing and session management, no system is 100% secure. You assume the risk of storing third-party credentials with us and should revoke GitHub app access from your GitHub settings if you have concerns.

10. Limitation of Liability

In no event shall the authors, maintainers, or contributors be liable for any claim, damages, or other liability arising from the use of the Service or generated content. This includes but is not limited to: inaccurate design tokens, incomplete extraction, service downtime, AI hallucinations, misuse of generated DESIGN.md files, or unauthorized access to stored OAuth tokens.

11. Disclaimer of Warranty

THE SERVICE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. Algorithmic inference of design tokens is approximate; human verification is recommended for production use.

12. Open Source & Self-Hosting

This project is open-source. The source code is available on GitHub. You are free to fork, modify, and self-host the application under the terms of its license. Self-hosted instances are your sole responsibility; these Terms apply only to the public instance we operate.

13. Modifications to Terms

We reserve the right to modify these terms at any time. Continued use of the Service after changes constitutes acceptance of the new terms.