Our principles
Seven rules we use to decide how a HiveCode tool should behave. They are descriptive — what we already do — and prescriptive — what we promise to keep doing.
- 01
The tool works first
Whatever you came for — the input, the output, the answer — has to be reachable without dismissing a banner, accepting cookies, signing in, or scrolling past advertising. We can monetise around the tool; we do not get to put anything in front of it.
- 02
Static is fast
Every page is pre-rendered at build time and served as plain HTML. There is no hydration step, no client-side router, no framework runtime. The first paint is the final paint.
- 03
No accounts, no tracking by default
We do not collect personal data on our servers, do not use first-party tracking cookies, and do not embed analytics scripts on the landing page. Where advertising is used on tool subdomains, it is disclosed on that subdomain’s privacy page and gated by consent in the regions that require it.
- 04
Bilingual by design
Both English and Spanish are first-class on every tool from day one. Strings live in a shared source of truth, hreflang and canonical tags are emitted on every page, and translations are written by humans on the team — not run through a machine translator and shipped.
- 05
Open when it makes sense
Most HiveCode repositories are public on GitHub. We do not publish things that contain secrets, vendor-locked clients, or work in progress that would only confuse readers — but if a tool is shipped, the code that built it is probably visible.
- 06
Accessible to the standard, not as theater
We target WCAG 2.2 AA on the user-facing surface: real keyboard navigation, visible focus, ARIA only where semantic HTML does not already do the job, sufficient contrast on real screens, prefers-reduced-motion respected. If you cannot use a HiveCode tool with a screen reader or keyboard, that is a bug, not a roadmap item.
- 07
Finish small things
We would rather publish five small, finished tools than one large unfinished platform. Scope creep is the failure mode we watch for hardest. If a new idea does not fit inside the existing tool, it becomes a separate tool at a separate URL — or it does not get built at all.