apotropic

Safety and security for AI ingestion

Turn away harm before it reaches the model.

Apotropic is the safety and security layer for data entering AI systems. Every crawl, retrieval, and upload passes a check that runs on your infrastructure, explains its decision, and honors the people whose content you collect.

See our products Get in touch
apotropic From apotropaic: intended to turn away harm, from the Greek apotrepein, "to turn away". Like the false eye on a caterpillar that makes a predator think twice.

How we build

Three rules

Data that enters a model is hard to take back out. We make the checks that happen before it gets there, and we hold them to three rules.

  1. Runs where your data lives

    Our engines run inside your infrastructure. Pages, documents, and results stay with you; nothing is sent to us to be checked.

  2. Every decision has a reason

    Deterministic rules you can read and test. Each verdict comes with stable reason codes and the evidence behind it, ready for an audit.

  3. Refusal comes first

    When a site says no, our tools stop. They detect traps so a crawler can back away, never so it can disguise itself or route around a refusal.

Our first product

One so far

CrawlSign

Pre-release / design partners

Crawler safety for AI ingestion.

CrawlSign runs inside your crawler and checks every page before it enters a dataset, an index, or an agent's context. It honors opt-outs, catches poison sources, tarpits, and hidden-link traps, and returns one decision with an audit trail.

Visit CrawlSign Product details 
{
  "url": "https://example.com/noai",
  "action": "stop",
  "score": 115,
  "reasons": [
    "ai_refusal_signal_detected",
    "meta_robots_disallow"
  ]
}
One decision per page. Output from the CrawlSign v0.1.0 CLI.

Building a crawler, a RAG pipeline, or an index?

We are working with a small group of design partners. If bad data getting into your systems is a problem you have, we would like to hear about it.

adi@apotropic.com