Don't be loved; be lovely.
That's the founding line. It's a reminder that the goal isn't to be popular or growth-maxing — it's to be genuinely good at the thing: making software that handles personal data with honesty and restraint.
The why
I started shenas because I kept running into the same frustration: tools that claim to help you understand yourself quietly make their business from the data you generate. The incentives are backwards. The thing that's supposed to serve you ends up serving advertisers, acquirers, or a future acqui-hire. Your health records, your sleep patterns, your finances — they leave your hands the moment you install the app.
The fix isn't to opt out of quantified-self software. The fix is to build it differently: raw data stays on your own hardware, the pipeline runs locally, and nothing leaves the mesh unless you explicitly ask it to.
What I believe
Privacy first, then wisdom, then science, then fun — in that order. Don't lie. Don't cheat. Don't moralize; be moral. Respect differences in goals. Say what you mean.
In engineering terms: privacy by default, raw data stays local, no telemetry. No rolling your own crypto. No moving records off-device. When aggregate insights are useful, they travel via differential privacy and secure aggregation — wisdom, not records.
The product
Shenas is a local-first quantified-self platform. It gathers health, finance, and lifestyle data from services you already use, normalizes it across your own home mesh — laptop, phone, tablet, server — and helps you find patterns. It is not a cloud service and is not trying to become one.
shenas is open-core: the community version is free, open source, and will stay that way. A commercial layer with hosted services funds the work. Anything that touches your data stays in the community version.
How it's run
Bootstrapped, no outside investment. Day-to-day operations are agent-led — roles like CTO, CISO, Legal, and engineering run as configured AI agents under founder oversight, working against versioned values and per-role briefs. The company runs on the same data pipeline the product offers users. If the architecture can't run a company, it probably can't run a person either — and I'd rather find that out on myself first.
Where things are
Stage 0: agent-first operations live, desktop alpha in daily founder use, agentic feedback system being dogfooded against company operations. The first public release will be a free, open-source community version for the quantified-self cohort, distributed via package repositories and word of mouth.