Launching today

Semdex
Your website, served as clean agent-readable documents
1 follower
Your website, served as clean agent-readable documents
1 follower
AI agents re-render and re-scrape your site on every visit. Semdex pre-compiles it into a semantic replica: clean HTML/Markdown/chunks, strong ETags, content negotiation, Content-Signal headers. Your site stays the source of truth.




Free Options
Launch Team

Render Sandboxes: a computer for every agentIsolated compute, persistent state and a real filesystem
Promoted
I built Semdex after watching agent traffic grow on my SaaS: every bot re-executes the SSR + DB + scraping pipeline, and every agent cleans the HTML differently. It felt like serving the same static document through a dynamic render path, once per crawler, forever.
Semdex moves the work to write time: when your content changes, we compile a semantic copy — cleaned, chunked, versioned, signed — and serve reads as static, cache-friendly responses. Content negotiation, 304s, progressive chunk reads, /llms.txt discovery, Content-Signal headers: the plumbing a serious agent endpoint needs, without rewriting your site.
First tenant: my own product site (contextlm.top → agent.contextlm.top). The /compare page shows exactly what changes. I'm not claiming GEO lift numbers yet — one honest measurement at a time. There's a free 7-day trial with 1M requests; paid tiers start at ¥19.9/mo. Happy to answer anything.