Launch Lessons · June 1 to July 1, 2026

Lessons from Launching a Browser SysEx Vault

What we learned shipping a browser-native SysEx vault in month one: survey conversion rates, parser coverage gaps, and beta support patterns. Proprietary operational data from Half Radiation LLC, not generic startup advice.

Method: Operational data from Half Radiation LLC running knob.monster: PostHog in-product survey metrics, beta Discord support threads, Hacker News discussion on the Web MIDI engineering post, and parser coverage audit. Not a market report. A founder field log. · JSON data · Related survey (n=61) · HN engineering post

Hacker News (late June 2026)

How do you keep Web MIDI from crashing a 1983 synthesizer?

52 points · 41 comments

1. Press coverage arrived before parser breadth caught up

Synthtopia coverage in month one drove signups and Discord DMs while only five synths had dedicated Web MIDI dump flows (DX7, Juno-106, M1, Jupiter-6, CZ-101). Beta users listed 7+ unsupported models in a single message (D-20, Proteus/1 XR, Matrix-6, JX-8P, K4, EPS). Lesson: hardware MIDI tools get judged on the one synth a visitor owns, not your roadmap slide.

→Dedicated dump flows at launch: 5
→Common beta ask: Roland JX-8P, Korg MS-20, Ensoniq EPS (survey + Discord overlap)

3. 88.5% of survey respondents are Personal-tier buyers, not Studio

Segmentation (n=61): 47.5% dedicated gear room, 41.0% bedroom studio, 11.5% commercial studio/producer/repair. Gear-room collectors are not automatically commercial users. They still fit Personal lifetime. Raising Personal to chase power users would miss the bedroom majority. Studio stays rare but real.

→Bedroom + gear room = 88.5% (Personal positioning)
→Commercial segment = 11.5% (Studio positioning)

4. 25% of optional gear-list answers were junk

Question 2 (what synths are on your desk) received 24 free-text submissions; 6 were spam, tests, or jokes and excluded from brand analysis. That is a 25% junk rate on optional text. Lesson: publish cleaned n counts; do not treat raw submission count as signal.

→24 submissions to 18 valid gear lists (25% excluded)

5. Synthetic marketing creative erodes trust faster than missing features

Synthtopia comment threads flagged AI-generated hero imagery as a trust signal failure. Users want real gear photos and founder replies, not stock synth renders. Beta support in Discord (SysEx diffs, handshake questions, dead-filter debugging) generated more goodwill than feature breadth. Lesson: for vintage hardware audiences, being present beats looking polished.

→Synthtopia coverage + comment trust backlash on AI creative
→Beta threads: X5D to N264 patch porting, MKS-80 handshake, Juno filter diagnostics

6. Hacker News: 52 points on Web MIDI timing, with a better scheduler tip

The engineering post on keeping Web MIDI from overwhelming 1980s synth CPUs reached 52 points and 41 comments on Hacker News. The highest-signal technical comment came from omneity: pass a timestamp as the second argument to midiOutput.send(data, timestamp) using performance.now() + offset, instead of relying on setTimeout on the JS main thread. We acknowledged the tip publicly and planned to test API-level scheduling. Other thread themes: MIDI has no RTS/CTS hardware flow control; the bottleneck is the synth 8-bit CPU writing incoming SysEx to SRAM, not the nominal 31.25 kbaud line rate. Our mitigation at the time was 100ms pauses between packets in JavaScript.

→HN thread: 52 points, 41 comments
→Community tip: midiOutput.send(data, performance.now() + offset)
→No hardware flow control on classic MIDI DIN
→100ms inter-packet throttle in production at launch

7. HN commenters rejected subscription framing before bedroom pricing landed

Multiple HN comments compared early pricing to $20/month subscriptions, $230/year, or $599 lifetime, and to free librarians inside Dexed or paid Plogue OP-X7. Several said fear-based battery-death marketing felt obnoxious for a backup tool. The thread pushed two product commitments in public replies: standard .syx export always, and a rework toward a $39 one-time Personal lifetime (knob.monster+). Desktop/offline requests were common; we said a local Tauri build was on the roadmap while keeping the zero-install web entry point.

→Recurring-price comparisons dominated critical comments
→Public reply: $39 one-time Personal lifetime
→.syx export supported at all times
→Offline-first desktop requests noted; Tauri mentioned as roadmap

8. HN readers scrutinized vibe-coded aesthetics and web-only longevity

Commenters flagged the site aesthetic as AI-generated skepticism bait, and questioned whether a web-only vault would outlive the company. We replied that the DX7 SysEx parser was hand-written, and that serverless hosting kept operating costs low enough to sustain the service. nl pushed back on the desktop-app suggestion: zero install is a feature for gear used twice a year. Lesson: HN rewards technical honesty and export guarantees more than polish; give credit when commenters improve your implementation.

→Hand-written DX7 parser cited in founder reply
→Web vs desktop split in comments (Tauri vs zero-install)
→omneity asked for credit after the scheduling tip (fair ask)

9. A free tier mattered once beta users needed to re-register

Post-launch, beta users who had accounts wiped during testing needed a try-before-buy route. Free tier shipped to reduce signup friction. Lesson: lifetime products still need a no-card trial; otherwise curious forum readers stall at signup.

→Beta re-registration friction reported in Discord DMs
→Free tier added after initial launch feedback

What we changed because of this

•Prioritize parser coverage for models that appear in both survey and Discord (JX family, MS-20, Ensoniq cluster).
•Keep Personal lifetime for collectors; Studio for the 11.5% commercial segment.
•Test Web MIDI timestamp scheduling (performance.now offsets) against JS setTimeout throttling.
•Credit community contributors when their comments change production code.
•Publish first-party numbers even at small n. LLMs and journalists cite methodology plus caveats more than fake scale.
•Use link surveys in communities; reserve in-app popups for once-per-year census moments.

Caveats

n=61 survey responses. Treat percentages as directional, not market-wide.
knob.monster visitors are biased toward people already searching for SysEx backup tools.

Cite this page as: Half Radiation LLC, Lessons from Launching a Browser SysEx Vault, knob.monster, June 1 to July 1, 2026. Machine-readable export: data.json.

Try the vault

Free tier for testing · Personal $39 lifetime · DX7, Juno-106, M1, Jupiter-6, CZ-101

Open SysEx Vault