What is INJECT.md used for?

INJECT.md audits LLM apps for prompt injection by running a battery of 46 payloads across five attack classes and returning a scored report. It is a good fit for testing an AI assistant that reads email or web pages and giving a security lead a scored injection report.

Yes. INJECT.md has a free plan, and paid plans start at $39 per month.

Is INJECT.md open source?

No. INJECT.md is proprietary (closed-source) software and can't be self-hosted. In the Security category, open-source options include ModSecurity, DefectDojo and CISO Assistant.

What are some alternatives to INJECT.md?

Other SaaS products in the Security category include Agnostics, aipwn and 42Crunch.

INJECT.md: features, pricing and alternatives

About INJECT.md

INJECT.md is a prompt-injection audit for applications that let a model read untrusted content such as web pages, emails, PDFs and tool outputs. Its battery has 46 payloads across five classes: direct override, obfuscation and encoding, indirect injection via ingested content, exfiltration side-channels and agentic tool abuse, and each payload cites the technique behind it.

The scored report gives a resistance score, a per-class breakdown and a verdict for every payload as self-contained HTML plus JSON for pipelines, and results that a string match cannot settle are marked as needing review. The site argues that system prompts cannot stop injection and that structural measures such as quarantining untrusted content hold up, and it provides a sandboxed live demo where you can try to leak a throwaway secret.

An audit of your endpoint is priced at 69 dollars, and 8 of the payloads are free to run. It is aimed at teams shipping LLM features that read external content and want a measurable baseline for how well their architecture resists injection.

Key features

46 prompt injection payloads in five classes
Scored resistance report with per-class breakdown
HTML report plus JSON for pipelines
Needs-review flags for unclear results
Free subset of 8 payloads
Sandboxed attack demonstration

Good fit for

Testing an AI assistant that reads email or web pages
Giving a security lead a scored injection report

Tags: prompt-injection; llm-security; owasp; ai-security; audit; testing

INJECT.md: questions and answers

What is INJECT.md used for?: INJECT.md audits LLM apps for prompt injection by running a battery of 46 payloads across five attack classes and returning a scored report. It is a good fit for testing an AI assistant that reads email or web pages and giving a security lead a scored injection report.
Is INJECT.md free?: Yes. INJECT.md has a free plan, and paid plans start at $39 per month.
Is INJECT.md open source?: No. INJECT.md is proprietary (closed-source) software and can't be self-hosted. In the Security category, open-source options include ModSecurity, DefectDojo and CISO Assistant.
What are some alternatives to INJECT.md?: Other SaaS products in the Security category include Agnostics, aipwn and 42Crunch.