Source code
You don't have to take our word for how nenva treats your data.
The code is public, and you can read it.
nenva makes strong promises: your data lives on your Mac, AI runs where you point it, and nothing about you is collected or sold. Promises like that are easy to type. The honest way to back them is to publish the source, so anyone who cares to look can check what the software actually does. nenva is two pieces of software, and both are on GitHub:
The desktop app
github.com/Anjadhe/nenva
Everything the app does on your Mac is in this repo: the apps, the assistant, email analysis, sync, storage. The installer you download is built, signed, and published from this repo by an automated pipeline you can watch run, so the code you can read is the code that ships.
The repo receives one snapshot per release rather than day-to-day development, so its history is one commit per version.
The hosted services
github.com/Anjadhe/anjadhe-connect
The optional services at api.anjadhe.com: web search for the assistant, nenva Cloud (AI for Macs that can't run a capable model locally), the encrypted mobile-sync relay, opt-in analytics, and feedback.
This repo is public for one reason above all: its core privacy claim is written in the code. Search queries and AI conversations are never logged and never stored; the database holds usage counters, not content.
Both repos are source-available under the PolyForm Noncommercial license: you can read the code, build it yourself, and change it for your own noncommercial use.