Micro Club / Independent research / September 2026
PointCast Micro Club V2
Product requirements for a home for Micro owners
Product, design, engineering, QA and community operations. Research snapshot: September 29, 2026, America/Los_Angeles. All target requirements, thresholds and commercial plans are proposals; local prototype scope is explicitly identified.
Build a small, reliable destination that makes a Micro immediately satisfying: get the controls working, play a distinctive short game, then keep or share something worth returning to. “Number one destination” is an ambition, not an evidenced market position.
Problem and evidence
Owners can already remap controls: Work Louder documents six layers and app-linked switching; Elgato and Logitech also maintain customization ecosystems [H02][H08][H10]. PointCast must add a compelling play collection and trustworthy setup evidence.
Codex Micro is currently listed as sold out [H01]. Compatibility questions appear in individual owner reports [H12][E05], which are useful design signals but not prevalence data. Start with existing owners and preserve ordinary keyboard/touch access.
Jobs and target sequence
Primary: “Help me make this device do something satisfying in three minutes.”
Secondary: “Show exactly which setup works, and let me restore mine.”
Returning player: “Give me a short new challenge and a little collection.”
Maker: “Give me a clear control grammar, a template and credit.”
Scope and status
Local V2 prototype scope
Production target - proposed
Three new games; seeded replay/challenge URLs; local personal bests
Existing PointCast authentication after repository audit; optional account after play
Mapped keyboard and touch controls; physical HID/RGB unproven
Versioned compatibility recipes, physical device QA and support ownership
V1 Signal Run, Pocket Echo and Bloom Room preserved
Saved profiles, challenges and curated presets with reporting/deletion
Deliberate boundaries
No mandatory wallet, hardware purchase, public feed, direct messages, creator payout marketplace, native HID bridge or LED-control promise in this release. Do not present a local passport as registration or a local best as a global rank.
02 / Game specifications
Three games, three tactile reasons to play
The V2 collection uses discrete semantic actions so a dial, joystick, ordinary keyboard or touch surface can reach the same game rules. Seeds reproduce content; they do not prove score integrity.
Dial In - precision
A 60-second run with five frequency locks. Previous/next actions move the tuner; the action key attempts a lock. A Micro dial maps to previous/next, with buttons and keyboard fallback.
Acceptance: exactly five locks are possible; completion or the 60-second deadline ends scoring once.
Show current frequency, target feedback, remaining time and lock progress; feedback works muted.
Same seed and game version reproduce lock targets. Restart clears timers and round state.
Switchboard - deliberate puzzle play
Solve three 4×4 pipe boards. Joystick directions select a tile; rotate changes its orientation. Hints deduct 70 points. Touch selection and rotate controls remain available.
Acceptance: each seeded board is solvable and advances only when a continuous valid connection reaches the destination.
Selection stays inside the 4×4 board; one rotate action makes one rotation. Show the 70-point cost before using a hint.
After the third solved board, show a final result and replay/share controls. No simultaneous-key requirement.
Pulse Club - rhythm
A 30-second, four-pad procedural rhythm round. Provide latency correction from -150 to +150 ms, explicit sound start/mute controls and visual timing cues.
Acceptance: one note can be scored at most once; held-key repeat cannot generate extra hits.
Apply latency correction consistently to timing judgments and keep calibration visible.
A seed reproduces the note sequence; after 30 seconds scoring stops and a result appears. Reduced motion preserves readable timing.
Shared acceptance
All three support start, restart, clear end state, local best and seed-sharing. Label scores “local” or “unverified.” Share links carry game/version/seed, never account secrets. Preserve V1 navigation and saved data; a failed storage write must not prevent play.
03 / Priorities and acceptance criteria
Reliability before accounts; identity before a feed
P0 is the required release behavior. P1 follows successful physical-device and return-play pilots. This page specifies target acceptance; it is not a claim that the public backend already exists.
P0 / Play and onboarding
Guest can start any game without registration. Offer save/account benefits after a completed round; dismissing the offer returns to play.
The key tester confirms every mapped semantic action. Setup explicitly distinguishes Codex Micro, Creator Micro 2 and ordinary keyboard.
Pause on focus loss/hidden tab, clear held controls and require a deliberate resume. Never capture gameplay keys while a text input is focused.
Keyboard-only users reach every interactive control with visible focus. Touch layouts avoid horizontal overflow at 390 px; mute and reduced motion remain usable.
Malformed challenge parameters use a safe fallback and visible explanation. Same seed/version yields the same game content across supported inputs.
P0 / Public-launch prerequisites - proposed
Audit existing PointCast auth, schema and routes before implementation; reuse the established identity system. No mandatory wallet or proof-of-purchase upload.
Ownership labels are self-declared unless separately verified; never imply vendor endorsement.
Every read/write enforces server-side authorization. Test another account cannot edit/delete a profile, preset or private run.
Client-submitted results remain unverified. A competitive leaderboard requires server-issued run context and independent validation before a verified classification.
Before public contributions: report control, moderation queue, rate limits, removal process, account/content deletion and a published privacy explanation.
P1 / Retention and contribution
Profiles can save collections and challenge history; guests can still play.
Curated presets carry author, device/OS/software versions, last test date and restore instructions.
Creators receive a small template, semantic input API and asset/accessibility checklist after first-party games pass QA.
A challenge opens the intended game and seed; recipient registration is optional. Public feeds, messaging and payouts remain out of scope.
04 / Proposed production design
A small backend with explicit trust boundaries
These data and endpoint contracts are proposed, not implemented. Confirm names and integration points in the PointCast repository audit; extend existing services rather than introducing a parallel identity stack.
Minimal data model
Record
Essential fields
Profile
Existing user ID; handle; visibility; self-declared device; created/deleted timestamps
Run
ID; user or anonymous session; game/version/seed; score summary; timing; validation class
Challenge
ID; game/version/seed; creator; optional public result; expiry/visibility
Preset
ID; author; versioned mappings; device/OS/Input versions; license; moderation status
Return server run ID, seed/version and expiry; rate-limit anonymous starts.
POST /micro/runs/:id/finish
Accept bounded result summary once; return validation class. Reject foreign, expired or duplicate runs.
GET/PATCH /micro/profile
Read permitted fields; authenticate writes; enforce handle rules and visibility.
POST/GET /micro/challenges
Create/read a bounded seed challenge; private reads require authorization.
POST /micro/presets and /micro/reports
Validate schema, authenticate submissions and queue public content for review.
DELETE /micro/profile
Authenticate, revoke access and apply the documented content/data deletion policy.
Integrity and privacy
Use validation classes local-only, submitted-unverified and server-verified. A seed or signed start token alone does not verify a score. Until a game-specific validator exists, exclude submitted results from competitive ranking.
Collect semantic product events and bounded score summaries, never raw keystrokes, OS input streams or device serial numbers. Any future replay validator must specify minimal semantic game actions and retention separately.
Proposed pilot retention: delete anonymous event-level data after 30 days; keep nonidentifying aggregates. Account-linked saves persist until deletion under the published policy. Document cookies/local storage and applicable consent behavior before public launch.
05 / Setup, quality and release gates
Make “Micro tested” an earned label
Vendor mapping guidance proves available configuration features [H02][H03][H04]. It does not prove these browser games work on every firmware, transport or browser. Publish tested combinations with dates, not a blanket compatibility claim.
Onboarding path
Choose device → select an unused game layer → map the common actions → test each control → complete one round → optionally save or share. Show how to switch back before changing mappings. Link vendor support for hardware/firmware issues; do not prescribe destructive resets.
Physical QA matrix - proposed
Dimension
Required coverage / evidence
Device
Codex Micro; Creator Micro 2 Base and Pro kept distinct; ordinary keyboard fallback
Transport
USB on each device; Bluetooth on supported models only
Desktop
Current stable macOS/Windows; Chromium plus Safari on Mac and Firefox; label combinations actually tested
Input/firmware
Record exact versions, app focus, layer and mapping; test with common monitoring conflicts present/absent
Controls
Dial both directions; four joystick directions; each pad; repeat/hold; blur/reconnect; restore original workflow
Fallback/accessibility
390 px touch, keyboard-only, muted, reduced motion, blocked local storage and malformed share link
Milestones and accountable roles
Milestone
Owner / exit evidence
M0 - local collection
Game engineer + designer: three game loops, preserved V1, reproducible seeds and browser QA
M1 - device pilot
QA lead + support editor: 12 observed owners; dated compatibility recipes and restore evidence
Product lead: pilot gate review; privacy text, support coverage and rollback plan
M4 - creator drop
Editorial lead: approved budget, license, QA and attribution for each contribution
Stop conditions
Do not claim hardware support if required controls fail or restoration is unreliable. Do not open public contributions without report/removal coverage. Hold release for cross-account data access, unrecoverable saves, scoring after completion or broken V1 routes.
06 / Measurement and commercial hypotheses
Use a small pilot to decide what deserves to grow
Thresholds are internal provisional rules, not industry benchmarks. Exclude team/test traffic; report recruitment and cohort sizes. Assess retention after all 30 activated owners complete follow-up through day 8.
Measurement definitions
Metric
Numerator / denominator / cohort
Activation
Unique setup starters completing a round ÷ unique setup starters; split by device and entry path
Setup success
Uncoached owners completing setup + one round in <3 min ÷ 12 observed owners; timer starts on setup entry
D7-window retention
Activated owners completing a session on days 6-8 ÷ 30 activated owners; first completion is day 0; cohort timezone recorded
Share utility
Unique recipients completing a shared seed ÷ 20 unique non-team recipients explicitly sent a challenge; also report landing-to-play conversion
Creator delivery
Invited creators delivering a QA-approved experience within budget ÷ 3 invited creators
Proposed go / no-go
Setup: at least 9/12 finish unaided in under three minutes; all 12 can restore their prior workflow. Otherwise repair onboarding before acquisition.
Return: at least 6/30 activated owners complete a session in the D7 window, days 6-8. Otherwise revise games or cadence before paid growth.
Sharing: at least 5/20 unique challenge recipients complete a round. Otherwise improve the invitation and recipient experience.
Creators: at least 2/3 deliver physical-QA-approved experiences within the agreed budget.
Release: zero known cross-account access defects; critical game/focus/save tests pass. Failed gates trigger interviews and fixes, not catalog or social expansion.
Commercial and resource hypotheses
Keep setup, tester and one game free. Test a finite $12 collection; separate interest, checkout starts and orders. Adjacent prices do not establish Micro demand [E07][E10].
Illustrative only: $12 less an assumed 15% variable deduction yields $10.20 contribution; a hypothetical $1,500 content/QA budget needs 148 sales, excluding tax and ongoing overhead. Replace assumptions with actuals before spending. Scope: three games, shared input, support and a 14-day pilot. Agree effort/rates separately; payouts, subscriptions and broad networking are deferred.