Skip to content
EarlyHound

Arcjet

1 comment
The first launch from Arcjet

Arcjet

AI agent runtime security

Developer toolsAI agents

Visit

Arcjet is runtime security that sits inside the agent path, not at the perimeter. It discovers which agents are running, then enforces policy on prompts, tool calls, and spend: prompt-injection detection, tool-call authorization, PII redaction, bots and abuse, plus an audit trail of what happened and why. Coding-agent hooks need no SDK; custom agents use guard(); HTTP apps use protect().

Meet the team

Sign in to comment

בן כהןEarly signal, tier 5Scout17h

Most agent security still lives at the proxy. Arcjet puts the decision next to the prompt and the tool call, which is where the damage actually starts.

Prompt injection, tool auth, and PII redaction in one runtime path is the rare combo that security and the people shipping agents both care about. I have not wired a Guard policy into a real agent yet; that is the first thing I would try.

What people who use it say

Nobody has reviewed Arcjet yet. If you have used it, you are the one who can say whether it holds up.

Write the first review