Built for the compliance moment Malaysia is in right now.

RedactNode started from a simple observation: employees at Malaysian hospitals, banks, and universities were already using ChatGPT and Gemini every day, whether or not their IT department had a policy for it. The question was never whether to allow AI at work — it was how to do it without MyKad numbers, patient records, or account details ending up on a server outside the country with no way to pull them back.

The 2026 PDPA amendments and the National AI Governance Guidelines (AIGE) made that question urgent for every DPO in the country. RedactNode is our answer: a proxy that sits between your staff and any external LLM, strips out regulated data before it leaves your network, and hands it back only once the response returns — so the AI still does its job, but the raw data never does.

What we actually built

Three pieces, working together: Tanda Guard, a browser extension that blocks PII pastes into public AI tools at the source; Tanda Proxy, the masking and de-anonymization gateway itself; and the Safe Zone Portalyou're looking at right now, a branded internal chat interface your staff can use instead. Every interception is logged as a one-way hash, never the underlying content — we built it so that even we can't read what got caught.

Where we're headed

We're working directly with early pilot organisations in healthcare, banking, and higher education to make sure the product matches how compliance teams actually work, not how we imagine it. If that's you, we'd like to hear from you.