> ## Documentation Index
> Fetch the complete documentation index at: https://firecrawl-docs-prompt-injection-screening.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Prompt Injection Screening

> Screen scraped page content for prompt-injection payloads before the content reaches your model. Enforced server-side.

Prompt injection screening examines the content of a scraped page for prompt-injection payloads. The screen runs before Firecrawl gives the content to you, and before your model reads it.

<Note>
  Prompt injection screening is an enterprise feature. It is available through your Firecrawl account team. Contact them to discuss it for your account.
</Note>

## Why scraped content is a risk

A scraped page is untrusted input. You control the request. You do not control the page.

Text on a page can contain instructions. Those instructions target the model that reads the page later — your agent, your extraction step, or your assistant. The page author writes them. Your pipeline delivers them.

Hidden text is the clearest case. A page can hide text in several ways:

* White text on a white background
* Text set to a font size of zero
* Text positioned off the visible area of the page
* Comparable tricks that keep text out of sight

A person who looks at the page never sees this text. A conversion to markdown or to plain text keeps it. The text then arrives in the context of your model as ordinary page content.

## What the screen does

* **Firecrawl classifies the content.** Scraped content goes through a content classifier before delivery.
* **Firecrawl flags suspected injection.** The classifier marks content that reads as an instruction to a downstream model.
* **The safe default is to withhold.** Firecrawl holds back flagged content. It does not pass the content through in silence.
* **Policy is configurable per account.** Your account team sets the policy that applies to your organization.

## This is not Lockdown Mode

Teams confuse these two features. They solve different problems, and one does not replace the other.

**Prompt injection screening inspects content.** It reads the scraped text and decides whether the text is safe to hand on.

**[Lockdown Mode](/features/lockdown) controls egress.** It serves the page from Firecrawl's index and cache. It never makes an outbound request to the target. It exists for compliance-constrained environments, where the request itself can leak sensitive information.

Two differences matter:

* Lockdown Mode does not inspect content. A cached page that carries a hidden instruction stays exactly as it was cached.
* Lockdown Mode changes how old a page can be. Under Lockdown Mode, an unset maximum-age default is much longer, so a served page can be older than you expect.

Choose by the risk you must control. Use Lockdown Mode when the outbound request is the risk. Use prompt injection screening when the page content is the risk. You can care about both.

## Availability

Prompt injection screening is available to enterprise accounts through your Firecrawl account team. The account team configures the policy for your organization and the handling of flagged content.

This page describes the behavior of the feature. It does not describe a request parameter or a response field. Talk to your account team about how the feature fits your workloads.