A Manifesto

Our Ethos.

The words we use, and what we mean when we use them.

Every company says it has values. Fewer companies define them. This page explains what we mean by the words we use, so you can hold us to them.

Human-First

AI should solve a person's problem before it showcases a technology.

Most AI companies lead with the technology. The model, the benchmark, the feature list. The person with the actual problem is an afterthought, and it shows: tools get built that impress other engineers but don't fit how real people work.

We start with the person. What are they trying to get done? What's eating their hours? What part of their week do they dread? The technology is only useful insofar as it answers those questions.

Human-first also means we never celebrate replacing a person. Much of the industry treats headcount reduction as a selling point. We think that's both wrong and shortsighted. The fear that AI exists to eliminate people is real, and companies that feed that fear are burning trust the whole industry depends on. Our work helps people get more done without adding extra hours. People stay at the center. That's the point of building any of this.

Eco-Conscious

We account for the footprint of the AI we build, and we act on it.

The dominant posture in AI right now is to scale as fast as possible and treat energy consumption as someone else's problem. Data centers grow, footprints grow, and the industry mostly shrugs.

We don't think using AI requires pretending it has no cost. A portion of every Orva subscription funds carbon removal through Stripe Climate, and each user can see an impact report showing what their subscription has contributed. We also engineer for efficiency: Orva uses smaller, more energy-efficient models wherever they can do the job, because the greenest computation is the one that never needed to run.

One rule governs all of this: the technology has to match the marketing. If we can't point to something concrete in how a product is built or funded, we don't claim it.

Explore Orva

Costly Signals

Trust is earned through commitments that cost something.

Values pages are free. Any company can publish one, and most do. That's why so few of them mean anything.

A commitment only proves something when it carries a cost. Contributing subscription revenue to carbon removal costs margin. Telling a prospective client to slow down and consider the security and infrastructural ramifications of what they're building costs momentum, and sometimes the sale. Choosing efficient architecture costs the easy shortcut. We make these choices because they're the only kind of evidence that separates a real position from a marketing position. When a company says it stands for something, look for the cost. If there isn't one, it's decoration.

Noticing

We keep track of how our work affects people and the world around us.

A lot of harm in technology comes from nobody paying attention. Systems ship, consequences accumulate, and everyone claims it wasn't their department. The faster the industry moves, the more convenient that blindness becomes.

We take the opposite posture. We pay attention to the energy our tools consume, to how automation changes what a workday feels like, and to what happens downstream of the systems we build. Paying attention doesn't make anyone perfect. It makes them accountable, and accountability is where every honest improvement starts.

Middleware for Humans

People shouldn't spend their day managing apps. Apps should be built well and handle rote tasks so people can ideate and innovate.

Modern work has quietly become app management. Copying information from one tool into another. Checking one dashboard against a second. Shepherding updates between systems that don't talk to each other. None of that is the job. It's overhead that grew until it swallowed the job.

The common answer is to add another tool, which adds another thing to manage. We build technology that sits in the middle and absorbs the overhead instead, so your time goes to judgment, relationships, and the work only a person can do. Success isn't how much you use our tools. It's how much less you have to think about tools at all.

Future-Focused

We build for the people who inherit the consequences of what gets built today.

The prevailing mindset in AI is quarterly. Ship now, scale now, deal with the fallout later, if ever. That mindset is how an industry ends up with security failures, public distrust, and tools people resent.

Future-focused means asking what today's choices compound into. Efficient models today mean a smaller footprint tomorrow. Honest positioning today means a trustworthy industry tomorrow. Helping a business modernize thoughtfully today means jobs and institutions that survive the transition instead of being casualties of it. The next generation will live with what this era of AI decides to become, and we build accordingly.

Ethical by Architecture

Principles belong in how systems are built, not just in what marketing campaigns say.

Right now, anyone can ship AI. People who have never considered security, data handling, or failure modes are wiring AI into businesses at record speed, and the wreckage is already visible: exposed records, broken workflows, systems that look finished but fall apart under real use.

Ethics that exist only as words are decoration. Real ones show up in structure: security treated as a fundamental rather than an afterthought, client data handled with care, honest scoping, and a willingness to say no when something shouldn't be built. Losing a project is acceptable. Shipping something indefensible is not.

Technology That's Invisible

Good tools increase human potential, not replace it.

The industry builds for engagement. More screens, more notifications, more time in-app, because time in-app is the metric that gets rewarded. The result is technology that demands attention instead of returning it.

We design toward the opposite standard: fewer interruptions, less to monitor, more finished work appearing where it should. Good technology is the kind you stop noticing because it simply works. We're not trying to be the center of your day. We're trying to give your day back to you.

In Closing

Why We Define Our Words

Vague values are how greenwashing happens. A company says "ethical," means nothing by it, and the word loses its meaning for everyone.

Defining terms publicly creates a standard that can be checked. If our actions ever stop matching these definitions, you'll have every right to say so.