QCue v1.1.1 is an incremental release on top of v1.1.0. Headline: voice capture is no longer tied to a single transcription provider — pick the speech-to-text (STT) engine you already hold a BYOK key for, or let QCue choose.
QCue itself is unchanged: a capture-first second brain — jot or speak a fleeting thought, a BYOK LLM distills your stream into a linked, Karpathy-style markdown wiki ([[wikilinks]], index.md, log.md), and a nightly Auto-Dream keeps it consolidated.
🤖 Works out of the box — the app ships pointing at QCue's hosted backend, so just sign up in-app (one tap with Google), no key needed to start.
The attached APK is upload-key (release) signed; on a sideloaded install allow "install unknown apps." Android 7.0+. Verify the download against the attached SHA256SUMS first.
📦 Downloads
| Android | iOS |
|---|---|
⬇️ arm64-v8a APK — 26.0 MB, virtually every phone from ~2017 on (Android 7.0+). Built from the public v1.1.1 tag. |
🍎 Download on the App Store — QCue: Capture Your Mind. iOS remains the functional reference (verified on iPhone 16). |
Verify integrity: the SHA-256 of the asset is in the attached
SHA256SUMS.Why one arm64-v8a APK? It covers essentially every phone from ~2017 on at ~26 MB — versus a ~64 MB "universal" APK that also bundles the emulator-only x86_64 and legacy 32-bit slices. (Need 32-bit? Build
armeabi-v7ayourself from thev1.1.1source.)
🎙️ What's new — multi-provider voice transcription
- 🗣️ Speak with any STT-capable BYOK provider. Mic transcription now runs on OpenAI, Groq, Zhipu, Gemini, or Qwen — whichever you've configured — instead of a single hard-wired engine.
- 🎚️ New Settings → "Voice transcription" picker. Choose Auto (QCue derives a sensible engine from the keys you've set) or pin a specific provider. There is no silent cross-provider fallback — your audio goes only where you chose.
- 🩺 Steadier transcription. Prefers a healthy STT key, auto-corrects the model name per provider, and gives clearer guidance when no STT-capable key is configured.
- 🐛 Fix: Qwen responses that omit the text part are now parsed correctly.
Full diff since the last release:
v1.1.0...v1.1.1· Changelog:CHANGELOG.md