Changelog
What changed in each ScribeKey release, newest first, with one page per release.
Every release is listed here and has its own page. These docs match ScribeKey 1.1.0 for Android.
1.1.0
Build 30, 17 November 2026.
What’s new
- New: Notes on your phone. Summaries, action items and meeting notes are written on your phone, with no account or key.
- New: Search by meaning. Find sessions by what they are about, not only the words in them, after an optional download.
- Send each finished session to Notion as its own page, through your Composio account.
- Setup asks for one download and one dictation, and the practice uses the real bubble.
- New skills: Flashcards, Key takeaways and Clean podcast.
- Fixes and improvements.
1.0.0
Build 29, 7 November 2026.
Closed testers got this release as build 28, on 7 Nov.
What’s new
- New: Flowcharts. Turn a session into a diagram in Studio, drawn on your phone and shared as an image or its source.
- New: Widgets and quick capture. Start a note or meeting from the home screen, a launcher shortcut or a Quick Settings tile.
- Setup is shorter: permissions come in two groups, and the Speech step recommends the live model for your phone and downloads it from the bottom button.
- Every speech model says when text appears, Live or After speaking, and Settings names your model once.
- Long dictation arrives in paragraphs, and learned words can join your Dictionary automatically.
- Fixes and improvements.
0.13.0
Build 27, 3 November 2026.
Not uploaded to Play. Testers get these changes first in 1.0.0 (29).
What’s new
- New: Flowcharts. Turn a session into a diagram in Studio, drawn on your phone and shared as an image or as its source. Mermaid diagrams in your documents and notes are drawn too.
- Open a diagram full screen to zoom in, or edit its source as code with a live preview.
- Add learned words automatically: turn it on in Dictionary settings or during setup, and the words you fix go straight into your Dictionary.
- Setting up a skills connection is three numbered steps. Paste a key and ScribeKey loads your models and tests the connection.
- Add tags right where they appear: type, press Enter, or pick one you already use.
- Fixes and improvements.
0.12.0
Build 26, 1 November 2026.
What’s new
- New: Studio. Turn a session into a summary, action items, an email or another document, then edit it with its own history.
- New: One app, one unlock. Try every feature for 7 days, then buy ScribeKey once to keep creating. ScribeKey Fast refines your dictation on your phone.
- Sessions open as one readable sheet, with notes, flags, types and tags, and documents you can edit, undo and back up.
- Dictation inserts at your cursor, keeps your first words while the speech model starts, and runs for up to 20 minutes.
- Fixes and improvements.
0.11.0
Build 25, 29 October 2026.
Not uploaded to Play. Testers get these changes first in 0.12.0 (26).
What’s new
- New: Documents. Run a skill on a session to make a new document, and your transcript stays as you said it.
- New: Rewrite with skills. After you dictate, pick a skill on the bubble and your words are rewritten in place.
- Dictate into any app for up to 20 minutes.
- Text already in a field stays put when you dictate into it.
- Fixes and improvements.
0.10.0
Build 21, 25 October 2026.
What’s new
- New: One app, one unlock. Try every feature for 7 days, then buy ScribeKey once to keep creating. Your Sessions stay yours either way.
- New: ScribeKey Fast. Refine your dictation with a model that runs on your phone.
- Sessions open for reading, and notes and flags edit in place.
- Cleanup keeps the numbers and names you said, and matches the app you’re writing in.
- Fixes and improvements.
0.9.14
Build 20, 22 October 2026.
What’s new
- New: Recently deleted. A deleted session waits 30 days, so you can restore it.
- Swipe a session to delete, copy or share it.
- Play a recording from its card, and jump from a search result to where it was said.
- Dictation no longer cuts off your first words.
- Fixes and improvements.
0.9.13
Build 19, 19 October 2026.
Not uploaded to Play. Testers get these changes first in 0.9.14 (20).
What’s new
- Type a note while Capture records.
- A simpler recorder, built around one large Pause button.
- Transcript view follows along as the recording plays.
- Works better at large text sizes, in split screen and with screen readers.
- Fixes and improvements.
0.9.12
Build 18, 17 October 2026.
Not uploaded to Play. Testers get these changes first in 0.9.14 (20).
What’s new
- New: Dictionary suggestions. ScribeKey notices your corrections and offers to learn them.
- Share a podcast episode link to transcribe it.
- Use Capture and Sessions with the bubble switched off.
- Fixes and improvements.
0.9.11
Build 17, 14 October 2026.
First build on Google Play, published to the closed test. QA testers received 0.9.11 as build 15 on 9 Oct; build 17 adds the Meetings, sharing and Settings changes made since.
What’s new
- New: Meetings. ScribeKey works out who spoke, and you can name each speaker.
- New: Skills. Rewrite your words with built-in skills or ones you import.
- Save sessions as Markdown files to a folder or your own WebDAV server.
- Share audio or video from another app to transcribe it.
- Fixes and improvements.
0.9.10
Build 14, 28 September 2026.
QA build, rebuilt with the same number on 28 Sep and twice on 4 Oct. This entry covers all four.
Added
- Add a speech model from Hugging Face without signing in, or from a file on this phone. Experimental and live models are listed for everyone
- Backups: include sessions in Android’s encrypted backup if you choose, back up daily to a folder you pick with a password, and move sessions to a new phone by device transfer. Settings shows when a backup needs you
- New setup: practise dictating in a chat, then take a short tour of Capture, the editor and Sessions
- Live words in Capture: with a live speech model, your words appear as you speak. Allow landscape is a setting in Appearance, off by default
Changed
- The accessibility disclosure now says when ScribeKey reads fields and what cloud rewriting sends. Battery setup opens Android’s battery optimisation list instead of asking for an exemption
- Also: choose Quill or Mumble for AI cleanup, with Standard cleanup the default during setup; Sessions can show Captures apart from app dictations
Removed
- Advanced Labs, its seven-tap unlock and the collapsible live transcript preview on Capture
Fixed
- Moonshine v2 no longer drops speech segments of 9.3 seconds or longer; an imported speech model no longer crashes setup; the editor no longer double-spaces lines; a saved Capture keeps its transcript and summary through failures and restarts
0.9.9
Build 13, 27 September 2026.
Added
- Dictionary words help Parakeet spell names and jargon
Changed
- AI cleanup models say what each one is best for
0.9.8
Build 12, 27 September 2026.
Changed
- Presets replace the Skills lab: one list under Settings › Enhance › Presets, with editable built-ins, import from a SKILL.md file, and typed instructions saved for reuse
- Cleanup samples show real output as one card, with removed and added words marked
Fixed
- Skills that were on no longer rewrite dictation after the lab is turned off; the side menu says speech is not set up when the chosen model is not ready
0.9.7
Build 11, 27 September 2026.
Same app as 0.9.6, renumbered for QA. No changes.
0.9.6
Build 10, 27 September 2026.
QA build. Covers everything since the last 0.9.5 QA build on 5 Sep.
Added
- Import a speech model from this phone, and use new on-device models including Qwen3, Omnilingual and Moonshine v2. The model list updates from a signed catalogue, with fallback download sources
- Cloud speech handles sessions up to 60 minutes
- Restore the original transcript after cleanup, and review every rewrite in one Improve sheet in the editor
Changed
- A 14-day trial: after it, without Pro, ScribeKey keeps recording but stops transcribing, and the bubble no longer appears in other apps
- Text sent to another app is confirmed before ScribeKey reports it delivered
Fixed
- The bubble stays above the keyboard; cloud speech no longer invents words in silence at the end of a recording; a failed connection test explains itself in plain words; times follow the phone’s 24-hour setting
0.9.5
20 August 2026.
QA candidate for the next Firebase beta cohort.
Added
- Current on-device Smart Cleanup path with Standard fallback and safer model lifecycle handling
Changed
- Refined Capture, Sessions, Dictionary, onboarding, and settings flows
- Hardened capture recovery, text insertion, session history, and QA diagnostics
0.9.4
13 August 2026.
QA build for the next round of beta testing.
Added
- A dedicated Capture workspace for recording, processing, and transcript stages
- A ScribeKey hub drawer for reaching the app’s main destinations and settings
- Field-aware rewriting from the transcript editor
- Recovery for paused and interrupted captures, including safer audio import handling
Changed
- Refreshed Material 3 Expressive colour, motion, navigation, Sessions, transcript details, and recording surfaces
- Model recommendations now account for device support and available storage
- Cloud credentials can be removed directly from settings
Fixed
- Capture finalization, audio import, overlay visualisation, and text insertion no longer lose state across their failure and focus transitions
- Rewriting readiness and consent races now resolve safely
- Grouped menus retain their native surface anatomy and accessibility navigation stays within the intended roots
0.9.1
3 August 2026.
Second beta build. Everything below landed after 0.9.0 was named, so testers who read the 0.9.0 notes have not seen any of it.
Added
- Downloaded speech models are checked against a published checksum before ScribeKey uses them
- About lists the open source projects ScribeKey is built on
- Cleanup formatting and endpointing are adjustable in Settings
Changed
- Icons are drawn from a self-managed Material Symbols set, and the two directional ones mirror in right-to-left layouts
- Release and QA builds are shrunk, so the app installs smaller
Fixed
- Restoring a backup no longer deletes your recordings, and a backup no longer carries your API key
- Upgrading from a build older than the database reset no longer crashes on launch
- The floating button keeps recording while you drag it
- Signing in to a cloud service completes when the browser returns to an app already running
- Leaving a screen now releases what it was holding, so long sessions stay responsive
- A cloud speech service that was never configured is refused instead of failing mid-dictation
0.9.0
2 August 2026.
First build handed to beta testers, through Firebase App Distribution.
Added
- Transcribe an audio file you already have: pick a file from Home, and the recording is kept, badged in history and retryable like any other
- The Dictionary now shows the corrections ScribeKey taught itself, and lets you delete one or clear them all
- Prompts can be pointed at the apps that use them, so email gets the email prompt
- Interrupted model downloads resume where they stopped, and free space is checked before one starts
- Failed and cancelled dictations stay in history with their audio, so you can retry them
- Spoken punctuation can be switched on, with a haptic to confirm it landed
Changed
- First run rebuilt around the deck pattern: one decision a screen, stated as its consequence
- The cloud screen now owns the connection, not the capability; each pipeline stage chooses its own placement
- Rewriting settled on one name, one writer and an amount that reads as a scale
- Every surface got an edge, and cards and buttons share one language
- Home leads with time transcribed, measured from real recordings
- The Settings hub lost its longer sections and now counts setup honestly
- Rewriting stages are skipped below a word threshold and abandoned on a timeout
- The app id moved from the legacy package ID to
app.scribekey - Added an About ScribeKey seven-tap unlock for Advanced Labs; experimental switches are managed in that single destination
- Gated HF Cloud, Audio Scribe, the Audio Scribe model-library route, its import control, benchmark tools, and stale onboarding selections behind the Labs feature gate
- Simplified normal Voice Engine and Enhance menus by removing one-option controls and redundant explanatory copy
- Updated Settings, About/Labs, Voice Engine, and Enhance wording for faster scanning
- Made Labs unlock-progress snackbars replace the current message instead of queuing older banners
- Updated M3 Expressive screenshot baselines for the reduced menu surfaces
Fixed
- Importing long audio no longer runs out of memory
- Transcripts no longer arrive with a leading space
- A custom prompt is no longer overridden by the system template, and its safety rules survive
- The dictation rate is weighted by how long each session ran, so one short fast burst no longer skews it
- Failed dictations no longer count towards time transcribed
- Cleanup strips only hesitation sounds, and honours the custom filler list
- Retry on Home actually retries, and a failure keeps you in the app
- Switching the cloud service away from OpenAI kept the OpenAI endpoint
- The overlay no longer blames the microphone for every failed dictation
- Stored HF Cloud or Audio Scribe selections now resolve to the safe local speech path while their Labs feature is unavailable
- Settings search and diagnostics no longer expose benchmark tools before the benchmark Lab is enabled
Removed
- The engine redesign prototype and its Labs entry
1.0.1
30 March 2026.
From before version numbers restarted at 0.9.0 in August 2026, so it is older than every 0.9.x entry above it.
Changed
- Timestamp grouping now shows only “Today” for today’s recordings; all older dates display as full locale-formatted dates with day name (e.g., “Saturday 28 March 2026”) — “Yesterday” label removed
- AI Enhancement screen now includes Benchmark Tools row (moved from main Settings > Processing card)
- Cloud AI engine option restored to selectable state — was incorrectly disabled with “Coming soon” label
- Voice Engine settings redesigned to mirror onboarding card layout; replaced piggy-bank icon with correct icon
- Auto-Cleanup / Smart Fix: removed master enable/disable toggle; all sub-features are always active
- Auto-Cleanup now exposes individual controls: Remove filler words, Auto-capitalize, Add punctuation, Custom filler words input, VAD silence threshold slider (500ms–3000ms), and Prewarm engine toggle
- Transforms section in Enhancement screen marked as “Experimental” with explanatory subtitle
Removed
- Benchmark Tools removed from main Settings > Processing card (now in AI Enhancement sub-screen)
- “Replacements” tab removed from Dictionary screen — Vocabulary and Shortcuts tabs remain
- Email Polish, Meeting Notes, and Chat Reply enhancement presets removed from the prompts database (migration 13→14 deletes these from existing installs)
- Starter prompts no longer seeded on fresh installs
Fixed
HistoryCardmodifier parameter reordered to be first optional parameter (lint compliance)grantedPermissionColor()composable renamed to start lowercase (ComposableNaming lint rule)Icons.Outlined.Savings(unresolved) replaced withIcons.Outlined.Paidin BackendOptionCards