Recursant, a mesh-based control plane for AI agents

Name: Recursant, a mesh-based control plane for AI agents
Availability: InStock
Author: hestefisk

by hestefisk·May 13, 2026·3 points·0 comments

Visit Project View on HN

AI Analysis

●●SolidBig BrainSolve My Problem

Istio-style sidecars for AI agents solving enterprise compliance gaps.

Strengths

•Sidecar pattern enforces network-level isolation without rewriting existing agent code.
•Hash-chained audit trails provide verifiable decision history for regulatory compliance.
•Supports multiple frameworks like LangChain and CrewAI via A2A protocol.

Weaknesses

•Zero stars and forks suggests unproven adoption and community support.
•Heavy dependency stack including Kafka and Redis increases operational overhead.

Post Description

I have submitted before, but it didn’t end up in the Show HN list for some reason.

I have built Recursant, an AI agent governance platform that governs agents across stacks and clouds to maintain compliance. In many large enterprises this is a real compliance issue with different teams using different frameworks and runtime environments, which creates control risk.

I hope to get some feedback on it. It is a viable alternative to the large vendor controlled stacks. Architecturally it is quite unique given its use of the Istio / sidecar pattern. This enables agent isolation at the network layer.

Thanks Anders