The BOM claimed "Zero W also works" as an aside. Checked it: Circle's
CHANGELOG gives gadget support for "Zero (2) (W)", the parentheses including
the plain Zero, and the README lists Raspberry Pi Zero as Tested with no
caveat. build.sh already emits kernel.img for RASPPI=1, and the MIDI gadget's
string descriptors are present in that image, so no build change is needed
for either board.
Keeping the Zero 2 W as the specified part on availability grounds. Records
that a plain Zero would permanently close the WiFi option in section 3a,
since it has no radio silicon at all.
Claude-Session: https://claude.ai/code/session_01TVCB25LBsmeteWvaSMz4Ne
Checked against the pinned Circle tree after the question of supporting
Bluetooth and WiFi on the Pi alongside USB.
Bluetooth is not possible on Circle, for two independent reasons: the stack
was removed for legal reasons and was classic Bluetooth in any case, with no
BLE or GATT anywhere in the tree, while MIDI over BLE is GATT; and the
remaining HCI transport would need a USB dongle in host mode, which the
gadget-only constraint already rules out.
That costs nothing. The WU-BT10 does GATT MIDI to the PC, and the PC is the
hub, so Bluetooth MIDI works today with no firmware change - it just
terminates one hop earlier. Section 3 now describes it that way rather than
as a spare part.
WiFi is possible via addon/wlan but is a project rather than a flag: Circle
lists WLAN as unknown on the Zero 2 W, ships no RTP-MIDI (only mDNS
advertisement of _apple-midi._udp), needs firmware blobs off the card at
boot, and its coexistence with gadget mode is unproven. Deferred, not
rejected; that last point is what to settle first.
Also marks Phase 1 done-but-unverified, and records that Phase 0 should bench
rpi_ws281x in SPI mode on GPIO10, since the stock examples use GPIO18 over
PWM and would prove the wrong wiring.
Claude-Session: https://claude.ai/code/session_01TVCB25LBsmeteWvaSMz4Ne
prosolis