Yew Search

Draft — being rewritten

Yew Search Principles

November 2026 · Alan James · Draft, not final

We own our tools and our data, and we build software that lets our users do the same. These principles apply to everyone who works on Yew Search: staff, contractors, and community contributors.

1. Privacy and self-hosting come first

Privacy and self-hosting are the first test for every decision, in the product and in the company. Yew Search promises users that their data never leaves their control, so we hold ourselves to the same standard.

  • We run our own services wherever it is practical, using Docker and Docker Compose on hardware we control or rent, such as Hetzner.
  • We avoid tools that track users, sell data, or need a cloud account to work.
  • We collect as little data as possible, about our users and about ourselves.
  • When a choice is close, we pick the option that keeps data in our hands.

2. Open software, with a narrow exception

We only use software that is open source or source-available, and that respects our right to repair: we can read it, fix it, fork it, or move away from it. Our own stack follows this rule, from Forgejo, Postgres, and Valkey to the Grafana stack.

A proprietary tool is a last resort, used only when no open alternative is good enough for real work. Every exception must meet three conditions:

  1. The reason is written down. We record which open alternatives we tried and why they fell short.
  2. We can leave. The tool must let us export all of our data in a usable format.
  3. It is reviewed every year. We check whether an open alternative has caught up, and switch when one has.

Exceptions are a reluctant trade-off, never a loophole. We keep them in a register like this:

Tool Why it is an exception Exit route Next review
Trello Self-hosted kanban boards were not sophisticated enough for our work. Plane to be evaluated before confirming. JSON export Yearly

Areas likely to need exceptions include payments, accounting, and payroll, where legal and banking requirements leave few open options.

3. Hardware we can open, fix, and keep

We truly own the machines we work on: we can open them, repair them, upgrade them, and run whatever software we choose.

  • Laptops: repairable machines such as the Framework, where batteries, memory, storage, and ports can be swapped by hand.
  • Operating systems: Linux by default.
  • Servers: commodity hardware, rented servers such as Hetzner, and Raspberry Pis, which Yew Search itself runs on.
  • Networking: open firmware such as OPNsense or OpenWrt rather than locked-down consumer routers, and coreboot where it is supported.
  • Office devices: nothing that needs a cloud account to work or constantly phones home.

The same exception applies here. Some work, such as building iOS apps, needs a Mac, and some hardware has no open option. In those cases we choose devices with good repairability, buy refurbished where we can, and keep them running for as long as possible.

4. Fair-code, with no bait and switch

Yew Search is fair-code from day one, and we tell everyone up front exactly how it will make money. Nobody who contributes to Yew Search or builds on it should ever feel misled.

  • Free for people, paid for businesses. Personal and homelab use is free. Businesses buy a commercial licence.
  • A public roadmap from the start. It shows which features will be part of the commercial tier, so people can decide with their eyes open.
  • What is free stays free. We add new features to the commercial tier. We never move existing free features behind it.
  • No surprise licence changes. We will not tighten the licence on code already released. HashiCorp's 2023 Terraform change is the example we refuse to repeat.
  • A clear contributor agreement. Contributors sign a short, plain-English agreement that explains their code may also be used in the commercial version, and why.

5. Small on purpose

We stay a small, tightly integrated team building one focused product. We would rather do excellent work at a sustainable pace than grow quickly and sloppily to chase a billion-dollar valuation. If great success comes, that is welcome, but it is not the goal.

  • We fund the company through grants, commercial licence revenue, and investors who share these values, and we stay profitable so we keep control.
  • We avoid funding that would push us towards growth for its own sake.
  • We hire carefully, for people who believe in this ethos.
  • Our self-hosted, open way of working is only practical because we are small, and that is part of the appeal.

6. How we treat people

In a small team built on shared values, how we treat people matters as much as the tools we choose. We admire businesses that stay small and profitable, but we borrow their business model, not their culture.

  • We are respectful and generous with teammates, users, and contributors.
  • We work in the open: decisions, roadmaps, and the reasons behind them are shared.
  • We credit and thank the people who help, inside and outside the company.
  • We answer the community directly and honestly, including when the answer is no.

7. Using these principles day to day

These principles are meant to be used, not framed. When you are unsure about a decision, ask these questions in order:

  1. Does it keep our data, and our users' data, under our control?
  2. Is there an open source or source-available option that is good enough?
  3. If not, have we written down why, confirmed we can export our data, and set a review date?
  4. Does it keep our promises to the community?
  5. Does it help us stay small, focused, and sustainable?

Anyone may question a decision using this document, and any exception to a principle must have its reason written down. These principles may be refined over time, but changes are made in the open and never break promises already made to our community.