What changed, release by release.

0.6.31 - Markdown (Obsidian) vault sync; assign speakers by name only; transcription reliability fix; library performance (faster search + listing). OpenVINO installs now move to the standard build automatically.

Swiss German experimental, onboarding restructure (1.5-2 GB first-run), model guard, white-screen fix

Swiss German routing fixes, live multilingual overhaul, CPU/dictation routing fix, dictation-after-stop, fast-segment debug controls.

What's new in 0.6.23

• Beta updates toggle — Settings → General → "Get beta updates". Opt in to receive pre-release builds earlier on a separate auto-update lane. Off by default; you can switch back anytime.
• Developer build-speed improvements (no change to app behaviour).

Installers

• directml — AMD / NVIDIA / general GPUs on Windows.
• openvino — Intel Core Ultra (NPU / iGPU).

The auto-updater picks the right variant for each install automatically.

LocalTranscript 0.6.22

• Fixes AI actions failing on machines with strict endpoint security (e.g. Cortex XDR): pinned AI runtime so it builds antivirus reputation, plus fast, clear errors and a re-download prompt if the model file gets quarantined.
• Fixed a layout glitch where the mic-only warning could shift the app sideways.
• Corrected the setup wizard's model download size totals.
• Retired the post-transcription referral popup (the referral program is unchanged, on the License page).

LocalTranscript 0.6.21

• Privacy + security hardening: Content-Security-Policy, self-hosted fonts, encrypted secrets at rest (DPAPI), a new 'Your data' panel, and safer file import/export.
• Diagnostics are now opt-in: a first-run consent step, off by default. Nothing leaves your device unless you choose to share anonymous diagnostics.
• Italian language support; Swiss German listed as coming soon.
• Free plan now shows Pro features (visible but locked); disk-space check before recording; setup wizard polish.

LocalTranscript 0.6.20

• Diagnostics reliability: clearer crash + performance reporting with quota-safe sampling (anonymous, EU-region, opt-out in Settings > Diagnostics > Privacy).
• Under-the-hood performance instrumentation. No change to how the app records or transcribes.

LocalTranscript 0.6.19

• New dictation-first home screen with a running time-saved stat.
• Anonymous, opt-out diagnostics (crash + performance) so we can fix issues faster: EU-region, scrubbed, never includes meeting content. Turn it off in Settings > Diagnostics > Privacy.
• Reliability and audio-capture robustness improvements.

0.6.18

Dictation is now the home screen

• Opens to Dictation by default ("speak, don't type") and it's the first sidebar item
• Active users see a "time saved by speaking vs typing" hero; new users get a short pitch + "Try dictation now"

Referral moved under Settings -> Licence

• The standalone "Refer" nav item is gone; referral now lives under the Licence tab
• Licence-aware enrolment (Pro users get invite framing instead of the free-Pro pitch; account email prefills from the licence)

Fix

• Record-start no longer fails when a previously-selected microphone is unavailable (unplugged headset / rotated id): it falls back to the default mic and validates devices against the live list (fixes the 0.6.17 "Start does nothing" regression)

0.6.17

Capture

• Per-input microphone and system-audio device pickers on the Capture page
• Live pre-recording input preview with per-input audio detection (Receiving / No signal)
• Live-transcript first-segment language correction

Performance

• Disabled the heavy local Outlook calendar scan (reverts to iCal): fixes slow Library/Dictation loads and the record-start delay

Reliability

• Recover speaker identification (diarization) on certain WebM recordings that previously failed at stop

New

• Referral program

New in 0.6.16

A clearer, consistent processing indicator

• While a meeting processes you now see one consistent progress bar with a time estimate across every stage: transcription, resolving speakers, and re-processing an imported recording. These used to look different from each other.

Pick any model for voice typing

• The dictation shortcut can now use any transcription model, including Large v3 Turbo. On a fast machine, choose the bigger model for higher accuracy (the smaller model stays the recommended default for instant response).

Better live transcription

• Fixed a filter that could drop phrases you repeated during a meeting, so the live transcript no longer loses legitimately repeated lines.

New in 0.6.15

Set up exactly what you want, faster

• Change each model (Transcription, Voice typing, and Summaries) right on the setup wizard's recommendation screen via a quick popup, instead of a separate "change models" detour.
• A new post-install step lets you set your dictation shortcut and try it on the spot: read a sample sentence in your meeting's language and watch it get typed back, with a live microphone check that flags audio that's too quiet, clipping, or noisy, so you know voice typing works before your first meeting.

Know your audio is being captured

• Live input preview before you hit record.
• A clear "Receiving / No signal" indicator next to each input so you can confirm your mic and system audio are coming through.

0.6.14 -- Cleaner playback and settings. Technical processing stats are now behind debug mode, with a plain-language Last run summary in the Analysis tab. The Live capture settings tab is renamed Dictation to match the sidebar. Removed the redundant Settings page header. Added a Free-plan usage overview on the License page.

What's new in 0.6.13

• Dictation is now a top-level item in the sidebar, opening a redesigned history grouped by date.
• Support menu in the sidebar — report a bug or send feedback in two clicks.
• Cleaner navigation — slimmer headers across the app and a tidier sidebar.
• Friendlier first-run setup — a clear "Recommended for your device" review step before anything downloads, with plain-language model names.
• Smarter model recommendations — high-quality Turbo for transcription, a fast model for live dictation.
• Polished launch splash — centered on screen with the brand slogan.
• Numerous capture, library, and settings refinements; full translations across English, French, German, Spanish, and Portuguese.

LocalTranscript 0.6.12

Reliability

• Live speaker identification no longer stalls mid-meeting. On long recordings a single stuck match could freeze all live name updates for minutes. Matching is now self-healing under transient processing hiccups (per-match timeout + drop-stale-backlog runner).

Privacy & onboarding

• The privacy/health indicator now correctly warns — amber dot and a red Privacy row — when transcription or live capture is routed to a cloud provider (it previously still reported "everything stays on this device").
• New welcoming Setup Wizard: a simpler first-run path that auto-picks hardware-recommended models, plus a "what language are your meetings usually in?" question. The manual model/cloud picker stays one click away.

Polish & localization

• Full UI translation parity across English, French, German, Spanish, and Portuguese (~1,000 newly translated strings).
• Plain-language Hardware settings; advanced/diagnostic knobs moved to an opt-in Debug tab.
• DPI-correct splash-screen centering at non-100% display scales, and dark-mode dropdown fixes.

What's new in 0.6.10

Connect a local Outlook calendar (new, opt-in)

Read your meetings straight from Outlook on your PC -- no iCal URL, no sign-in, and no network. It reads the calendar already synced on your machine (via classic Outlook or the Windows appointment store), so it keeps working behind corporate proxies and firewalls that block the iCal feed. Find it under Settings > Integrations > "Connect local Outlook calendar". Your existing iCal setup is unchanged.

Better iCal diagnostics

Clearer errors and detailed logging when an iCal feed fails (including a specific hint for corporate-proxy 5xx errors), plus bounded timeouts so a stalled feed fails fast instead of hanging.

Note: the local Outlook calendar is brand new -- if you hit any issue connecting it, the iCal URL option still works exactly as before.

Fixes live speaker identification (it was silently failing) and corrects the live transcript audio-source labels (system audio could be shown as Mic audio).

Separate models for live vs final transcription: pick a lighter/faster model for the real-time transcript while keeping a higher-accuracy model for the final pass at stop. Plus satisfaction controls that gate the stop-time re-transcription.

Maintenance release. Fixes 0.6.0, which shipped with a stale frontend bundle on some installs. Live speaker identification (opt-in) now works correctly: provisional speaker labels appear in real time while you record. Final speaker assignments are still produced by the full-audio diarization pass at recording stop. Disabled by default -- enable it in Settings.

Live speaker identification (opt-in): LocalTranscript can now show provisional speaker labels in real time while you record, instead of only after you stop. Final speaker assignments are still produced by the full-audio diarization pass at recording stop. Disabled by default -- enable it in Settings.