hoop.dev
Runtime controls for AI agents
hoop.dev is a security gateway that sits between AI agents and infrastructure, inspecting commands, masking sensitive data in responses, and routing risky actions for human approval. It deploys as a sidecar proxy with no code changes or new credentials, applying controls at the wire protocol layer across PostgreSQL, MySQL, kubectl, SSH, and gRPC.
Its main modules include Data Masking, Guardrails, Session Analyzer, and Human in the Loop. Guardrails block destructive commands, data masking hides PII and secrets from agents, the Session Analyzer scores actions for intent and syntax risk, and high-risk writes route to a human in Slack or Teams. The gateway is MIT-licensed and open source, with a control plane for Enterprise that manages rules across sidecars.
It is aimed at engineering and security teams governing access by engineers, AI agents, and service accounts. The open source edition is free and self-hosted, with per-sidecar limits on guardrails, masking rules, and Session Analyzer rules. Enterprise removes those limits, adds a control plane, support, and SLAs, and is priced per identity based on the environment.
12 alternatives to hoop.dev
Ranked by how well each tool replaces hoop.dev: shared features, audience, price and popularity.
The open-source alternative to Claude Managed Agents
Covers 7 of 15 key features.
Free planOpen source65 out of 100 matchFree- 61 out of 100 match$120/mo
Turn unstructured knowledge into real-time context for AI agents
Covers 5 of 15 key features.
Free plan61 out of 100 matchFreeThe AI agent platform for customer experience automation.
Covers 4 of 15 key features.
Free plan61 out of 100 matchUsage-basedConversational AI agents that work in the real world, built on the OASYS voice-native Orch
Covers 5 of 15 key features.
61 out of 100 match—Sovereign AI workspace without vendor lock-in.
Covers 7 of 15 key features.
Free planOpen source60 out of 100 match$10/mo- 59 out of 100 matchFree