Agyn: An Open-Source Platform for AI Agents with Scalable On-Demand Infrastructure

Forum topic · 小凯 · 2026-05-29

Summary

This paper introduces Agyn, an open-source platform by Nikita Benkovich and Vitalii Valkov (arXiv:2605.27575) designed to address the shift from building individual AI agents to operating them at scale in production. Production AI agents execute non-deterministic workflows, maintain stateful sessions, and often run with privileged access to internal services, requiring proper isolation, governance, and security. Agyn is built on three core principles tailored for agent workloads: a signal-driven, stateful serverless runtime on Kubernetes; a Terraform provider for defining agents and harnesses as infrastructure-as-code; and a security model grounded in zero-trust and least-privilege principles. The platform is agent-agnostic, model-agnostic, and cloud-agnostic, making it adaptable across frameworks and deployment environments.

Overview

Field: AI Authors: Nikita Benkovich, Vitalii Valkov Published: 2026-05-28 arXiv: 2605.27575

Abstract (full translation)

As organizations move toward production deployments of AI agents, which execute non-deterministic workflows, maintain stateful sessions, and often operate with privileged access to internal services, the engineering challenge shifts from building individual agents to operating them at scale with proper isolation, governance, and security. In this paper we present Agyn, an open-source platform designed around three key principles tailored for agent workloads:

A signal-driven, stateful serverless runtime on Kubernetes;
A Terraform provider for agent and harness definition;
A security model grounded in zero-trust and least-privilege principles.

Agyn is agent-agnostic, model-agnostic, and cloud-agnostic.

Key points

Focuses on the operational challenge of running AI agents at scale, not on building individual agents.
Serverless, stateful runtime on Kubernetes handles non-deterministic, session-based agent workloads.
Infrastructure-as-code (Terraform) for reproducible agent and harness deployment.
Zero-trust, least-privilege security for agents with privileged internal access.