Tork – Open-source AI governance layer, 11 SDKs& 116 framework adapters

Name: Tork – Open-source AI governance layer, 11 SDKs& 116 framework adapters
Availability: InStock
Author: yusufjacobs

by yusufjacobs·Feb 12, 2026·1 point·0 comments

Visit Project View on HN

AI Analysis

●●SolidNiche GemBig Brain

Policy-enforced AI agent middleware with 116 integrations, but governance-as-code is nascent.

Strengths

•~20ms latency overhead is impressively low for an interception layer
•Cryptographic hash-chained audit receipts prove compliance decisions for regulators
•50+ PII types across 13 countries + MCP-native design catches emerging agent patterns

Weaknesses

•AI agent governance is still early—unclear if orgs will adopt governance layers at scale
•No case studies or production user testimonials provided

Post Description

Been working on this solo for a while. Tork is a governance middleware that sits between AI agents and the tools they call — intercepting every interaction to apply PII redaction, policy enforcement, and compliance audit trails. It handles PII redaction, policy enforcement, and compliance audit trails. Works with 116 adapters across 11 programming languages. ~20ms latency overhead. Free tier available. https://tork.network Would love to hear thoughts from anyone dealing with AI governance challenges in production.