Show HN: Bayleaf – Building a low-profile wireless split keyboard

Original Article ↗ Hacker News Discussion ↗

Overall reaction & aesthetics

  • Strongly positive response to the build quality and styling; many say it looks like a polished commercial (even Apple-level) product.
  • Several people describe it as their “dream” or “grail” board and explicitly say they’d buy it, even at a premium.
  • Some admit it’s “not for me” ergonomically, but still praise it as a beautiful artifact and an impressive learning project.

Ergonomics & typing experience

  • Many users like split keyboards and low-profile switches for reduced wrist extension and joint strain, comparing it favorably to typical “low profile” mechs that are still quite tall.
  • Others strongly prefer concave keywells, columnar stagger, tenting, and/or palm rests (e.g., Kinesis, Glove80, Dygma, ZSA) and doubt they could go back to a flat, non‑curved surface.
  • Ortholinear gets mixed reviews: some say it’s more natural, easier on fingers, and good for layer-based layouts; others report pain or find staggered columns and vertical stagger more comfortable.
  • There’s a long subthread on correct keyboard tilt (negative/“downward” vs the common positive tilt) and wrist posture, with some disagreement but general emphasis on ergonomics being highly individual.

Layout, layers, and legends

  • Non‑standard, ortholinear 60% layout without printed legends is divisive:
    • Fans argue layers (numbers, arrows, nav, symbols) on home row and thumb keys are powerful, reduce hand travel, and quickly become muscle memory (e.g., Miryoku-style).
    • Critics miss dedicated arrow/home/end/pgup blocks and worry about learning curve and rare-symbol recall.
  • Multiple people explain that legended caps are expensive and layout‑specific; blank or minimally legended caps fit the niche, programmable nature of these boards.

Wireless, firmware, and electronics

  • The build uses nice!nanos and ZMK; commenters confirm ZMK supports multi‑device BT, NKRO, layers, and a “primary + peripheral” or dongle mode for splits.
  • Some praise nice!nano convenience (integrated battery management); others complain about the “keyboard tax” vs much cheaper generic nRF52 boards.
  • One detailed thread discusses whether each half could simply be an independent keyboard versus coordinated halves doing layer logic on-board.
  • A minority consider wireless a dealbreaker due to latency, reliability, and HID security.

Cases, materials, manufacturing & cost

  • The machined aluminum case is widely admired and seen as solving common DIY issues like flexy PCBs and loose switches.
  • Discussion compares 3D-printed (FDM vs SLA/MJF) vs machined metal cases, with cost and small-run economics cited as major constraints; steel would be heavier/quieter but pricier to machine.
  • The prototype’s BOM is about $400; with tools and software, total project spend is estimated around $1K+. This leads to debate about realistic retail pricing (likely well above the ~$150 some wish for).

Alternatives, features & wishlists

  • Many cross‑reference similar or alternative boards (Moonlander, Voyager, Glove80, UHK, Corne-ish Zen, etc.).
  • Strong interest in a future production run and/or DIY kits; people suggest:
    • Optional wired mode, trackpad/trackpoint or trackball integration, magnetic pogo‑pin charging and attaching, multi‑device switching, and possibly more thumb keys or columnar/vertical stagger.
  • Several note they’ve long wanted “a Magic Keyboard, but split in half”; some currently approximate this using two separate Magic Keyboards side by side.