What are the five levels of AI-agent friendliness measured by Silicon Friendly?
Silicon Friendly measures websites across five levels: L1 Basic Accessibility (Can agents read content?), L2 Discoverability (Can agents find things?), L3 Structured Interaction (Can agents talk to it?), L4 Agent Integration (Can agents do things?), and L5 Autonomous Operation (Can agents live on it?).
How many individual checks does Silicon Friendly perform on a website?
Silicon Friendly performs a total of 30 individual checks across the five levels to assess a website's AI-agent friendliness.
What does an L0 rating signify for a website?
An L0 rating, such as for www.blackbear-solutions.com, indicates that the website is 'not silicon friendly yet,' often due to protective measures like Vercel Security Checkpoints or bot-detection challenges that prevent AI agents from accessing its content. Does Silicon Friendly provide recommendations or tools to improve a website's score?
Silicon Friendly primarily functions as an assessment tool, providing a detailed breakdown of passed and failed checks. While it identifies areas for improvement, it does not offer direct tools or recommendations for remediation.
What kind of information is checked under the 'Structured Interaction' (L3) level?
The 'Structured Interaction' (L3) level checks for the presence of a structured REST or GraphQL API, consistent JSON responses, API support for search and filtering, an A2A agent card at /.well-known/agent.json, documented rate limits with proper 429 responses, and structured error responses with codes and messages.