Tip Of AI
Discover what you can do with AI
Arcjet is a runtime security platform designed to protect AI agents, large language model applications, coding agents, and web applications. It works inside application code through SDKs, agent hooks, OpenTelemetry integrations, and framework integrations rather than requiring a gateway, proxy, sidecar, or network migration. Arcjet captures agent actions and groups them into runs, providing visibility into users, sessions, routes, actors, tools, typed arguments, and previous steps. Its policy engine can allow, block, redact, or hold actions for review before execution. Security controls include prompt injection detection, agent tool authorization, MCP and API monitoring, sensitive information detection, data loss prevention, token and spending budgets, bot protection, WAF capabilities, rate limiting, email validation, signup protection, content moderation, and abuse prevention. Policies can be managed in code for version control and testing or remotely for immediate security changes. Arcjet supports JavaScript, Python, and Go, along with frameworks and platforms including Next.js, Express, FastAPI, LangChain, LangGraph, Claude Agent SDK, OpenAI Agents, Vercel AI SDK, Cursor, Claude Code, Codex, and Copilot. It reports less than 1 millisecond local decision overhead and approximately 20 to 30 milliseconds when using its cloud API. The platform supports SOC 2 Type II, custom retention policies, private storage options, SIEM streaming, and more than 600 bot types across 25 categories.