The Semantic CDN
for AI Agents
Crawl once, serve everyone. A shared cache of structured page representations that eliminates redundant web crawls.
Crawl once, serve everyone. A shared cache of structured page representations that eliminates redundant web crawls.
| URL | Title | HTML | SOM | Ratio | Crawl | When |
|---|---|---|---|---|---|---|
| Loading... | ||||||
Pages are categorized by freshness profile. Each category has its own TTL rules.
Any AI agent sends a GET request with the target URL. If the page is cached and fresh, the SOM representation is returned instantly.
On a miss, the cache fetches the page via a headless browser, converts it to SOM using Plasmate, and stores the result with a category-based TTL.
Subsequent requests for the same URL are served from cache. No redundant crawls, no wasted tokens, no duplicate browser sessions.