Stop burning agent token budgets on HTML noise.
Onto is an edge‑first infrastructure layer that sits in front of your React application. We extract semantic substance into compact Markdown payloads and serve those to AI agents via HTTP content negotiation—while preserving the full HTML experience for humans.
The modern web is optimized for humans,
not AI agents.
When AI crawlers ingest modern React apps, token budgets are burned on layout, tracking scripts, Tailwind classes, and animation wrappers.
A new class of crawlers and IDE agents explicitly request Accept: text/markdown. Without a dedicated negotiation layer, you risk hallucinated discounts, outdated feature descriptions, and missing citations.
Four layers of control.
Philosophically, Onto avoids creating a shadow web. We faithfully represent the same underlying content in multiple formats based on the consumer's ontology.
AIO Score Calculator
Public infrastructure to assess 'agent-friendliness' via HTML bloat vs semantic text analysis.
@onto/next Extractor
Traverses route trees, strips visual noise, and compiles semantic pages into pure Markdown/JSON.
Edge Middleware
Sub-15ms edge routing detecting GPTBot/Markdown headers to serve optimized payloads instantly.
Control Plane
Agent analytics and hidden prompt injection targeting AI logic without altering human UX.
The industry is moving to Markdown.
The core thesis of Onto is aggressively validated by recent research on LLM context windows and infrastructure moves by major hosting providers.
Vercel Content Negotiation
Vercel officially documents that serving Markdown instead of HTML to AI agents reduces payload size by ~99%, lowering latency and increasing ingestion success rates for models like GPT-4.
Anthropic "Lost in the Middle"
Research proves that "token noise" (HTML tags, generic CSS) degrades a model's ability to extract facts. Stripping the React Tax actively reduces hallucination rates in enterprise RAG.
Enterprise AI Crawling
Extracting data from strict enterprise domains forces AI companies to use expensive proxy arrays to bypass WAFs. Onto provides a cheap, compliant backdoor, saving millions in compute.
Standardize for the Agent Web.
Provision early access to @onto/next and the Edge Infrastructure core. We are onboarding technical docks and commercial storefronts first.