Custodyn — AI Agent Runtime Security Platform

Stop AI agents from going rogue — security embedded inside agent code.

by atuljetli

Visit

Aug 28, 2026 Security & Privacy

ai safety ai_agents developer-sdk human-in-the-loop runtime-security

About

A runtime security platform for AI agents. Developers install the Python or JS SDK directly into LangChain, CrewAI, AutoGen, or OpenAI Agents SDK pipelines. Every agent action is intercepted before execution — risky actions blocked, medium-risk routed to human approval, everything logged in a tamper-evident audit trail.Included in sale:Live platform at custodyn.appPython + JS SDKs on PyPI and npm (MIT licensed)9+ framework integrations: LangChain, LangGraph, CrewAI, AutoGen, MCP, Claude Code, OpenAI Agents SDKHuman approval queues, policy engine, multi-tenant dashboardSecurity pentest July 2026Open-core model ($49–$2000/mo tiers)60 day handover supportPre-revenue. Full acquisition. Motivated seller.