I What Anvyr Is
Anvyr is a software company. We build our own products and we build software to order.
When a client comes to us, we learn their domain first, then build the right thing from scratch. We don't reskin templates, we don't outsource, and we don't hand off half-finished work. The same standard applies whether the product has our name on it or someone else's.
We make software for people who read source code before they read landing pages. Our users don't want magic. They want leverage.
II Principles
01 Intention over convention
Every name, every architecture, every default is a deliberate choice. We do not cargo-cult patterns from other codebases. If a decision cannot be explained in one sentence, it hasn't been thought through enough.
02 Code is craft
Software is a creative discipline. Every function signature, every error message, every line break is a design decision. We treat them that way. The codebase is the product, not just the output of the product.
03 Speed is respect
Fast is respect — for the user's time, for the client's, for each other's. We do not ship loading spinners where instant response is possible. We do not hold a bug until Monday or an email until next week. Software that wastes cycles disrespects the user; a team that wastes time disrespects everyone waiting on it. Urgency is the default, not because we're reckless but because delay is a form of disrespect. Speed without precision is chaos. Precision without speed is irrelevance. We refuse both.
04 Ship in public
We do not polish in private. Work happens in the open — commits, decisions, mistakes — because building in public forces honesty. A feature that exists and works today beats a perfect feature that exists next quarter. But "ship fast" is never an excuse for shipping broken: everything that leaves the door must function correctly. We iterate in public, not in committee.
05 Manuals are a tax
Good software doesn't need a manual. Good architecture doesn't need a diagram. If the user has to read documentation to understand a basic workflow, the workflow is wrong. Every page of docs is a tax on the user's attention — we keep the bill small.
06 Own the stack, or own nothing
When we depend on something, we understand it completely. When we build something, we build it from the ground up if the alternatives don't meet our standard. We don't wrap bad abstractions in pretty APIs — we replace them.
III Structural Constraints
These are not guidelines. They are governance decisions, designed to survive growth, leadership changes, and external pressure.
Permanent Private Structure
Anvyr will never go public. No IPO, no SPAC, no direct listing. Ever. Public markets create legal pressure to maximize quarterly returns, which is incompatible with building software with intention. We fund growth through revenue and aligned capital. The company answers to its users and its craft, not to shareholders demanding exits.
Naming
Names are permanent. They carry identity. Anvyr names for direct products are constructed, phonetically sharp, and intentional. They are not acronyms, not descriptions, not inside jokes. A name should be pronounceable in any language and memorable after hearing it once.
The rule binds direct products. Sub-brands and acquired projects keep their existing names — VelvetCMS is the standing example. The discipline lives at the boundary; it does not extend past it.
Quality gate
Nothing ships without the author being willing to defend every line. "It works" is not a quality bar. "I'd build it this way again" is.
Source availability
Some Anvyr projects are open source. Others are not. We choose per-project, based on whether opening the source serves the work — not as a marketing posture either way. When a project is open, the code stands on its own and we accept the scrutiny that comes with it. When a project is closed, the same standard applies internally.
IV What Anvyr Is Not
We are not a body shop. We don't place developers at desks and bill hours. When we take on client work, we take ownership of the outcome.
We are not optimizing for growth metrics. Revenue matters because it funds the work. The work is the point.
We are not slow. If it looks like we're waiting, something is wrong.
V On Working With AI
AI is a tool in the workshop, not the carpenter, not the client. We use AI to amplify craft, not to replace judgment. Every AI-assisted output is reviewed, understood, and owned by a human. If we can't explain why the code is right, we don't ship it.
Build things that work,
for people who care.
This document is a living artifact. It evolves as Anvyr evolves, but the core never changes.