What we do with your code and data
This page lists what we have certified, what we have not, where your work lives, and which AI tools touch it. If anything here is unclear or out of date, write to [email protected].
Last updated October 12, 2026
Who we are
Our registration documents hang on the office wall, as Nepali law requires. If you are making the trip, you are welcome to visit and meet the people doing your work.
Certifications
| Standard | Status |
|---|---|
| SOC 2 | No report |
| ISO/IEC 27001 | Not certified |
| ISO/IEC 42001 | Not certified |
We would rather tell you this plainly than imply more. If any of it changes, it will show up in the change log at the bottom of this page.
Where your work lives
Today, client code and data stay in the client's own systems, and we work inside them. If an engagement needs us to host something on our side, where it goes is agreed in the contract before anything moves.
Services that touch client material
| Service | What we use it for |
|---|---|
| Cursor | Code editor, AI assistance, and coding agents, including agents running on xAI's Grok models |
| Anthropic (Claude) | AI assistance with code and documents |
| Zoho | Documents, contract signing, internal messaging |
How we use AI on your work
We use AI tools on client work, and you should know exactly which ones and under what terms.
- Cursor is our team's editor, and our Grok-based coding agents run inside it. Privacy Mode is enforced across our team by admin setting, so no one can switch it off. Under Cursor's terms, that means your code is not used for training, and Cursor holds zero data retention agreements with its model providers, including xAI. Cursor's privacy documentation
- Claude is used on an individual plan with the model training setting turned off. Under Anthropic's consumer terms, that means conversations are not used to train models. Anthropic still keeps them for a limited period and may review content its systems flag for safety. Anthropic's privacy policy
- Agents and data access. Our agents sometimes need to read a database to do their work. Each one gets the narrowest access the task allows, and where the work permits, we keep agents away from client data entirely.
Reporting a security issue
Write to [email protected]. A person reads it, and we will reply with what we found and what we are doing about it.
Documents
We are happy to sign an NDA before any detailed discussion. For anything else you need for a vendor review, ask, and we will tell you plainly what we have and what we don't.
Change log
October 12, 2026: Page published.