ENTERSPEAK / CHANGELOG

Every step.
Better words.

What we have added, improved and prepared for EnterSpeak. Follow the app and website changes, with delivery status kept clear.

Updated 2026-10-02

App updates

Android 0.1.0-rc11 · Build 21

Android release submitted to Google Play

  • Prepared the first Android release with on-device dictation, Cantonese writing styles, meeting notes and optional local speaker groups.
  • Includes an English keyboard and cursor controls, phrase-based translated captions, recording history, sharing and PDF export.
  • Bundled local AI models and phone/tablet layouts. This free release requires no EnterSpeak account. Submitted for review; public availability has not been confirmed.
#
Source update

Apple · source update

Language-model and speech-service compatibility improvements

  • Optimized language-model processing and improved compatibility checks and guidance for optional speech-recognition services.
  • Local recognition remains the default. This source update was made after the build 54 packages and is not included in those prepared installers.
#
Package prepared

Apple 1.0.0 · Build 54

Choose your own speech-recognition API

  • Added independent settings for optional speech-recognition and writing services on Apple platforms.
  • Improved speech-service compatibility. Explicit consent is required before audio is sent to the configured provider; provider fees are separate.
  • Signed iPhone/iPad and Mac packages are prepared for testing and upload. These records do not establish store publication or paid-provider accuracy.
#
Uploaded for testing

iOS 1.0.0 · Build 53

English typing and cursor controls

  • Added an English QWERTY keyboard with ABC/Voice switching, Shift, Caps Lock, numbers and symbols. The selected keyboard mode is remembered.
  • Added left/right cursor movement and up/down navigation across explicit newlines, plus a wider space key. Manual typing or cursor movement cancels pending voice insertion.
  • English typing does not require microphone access or Full Access. Uploaded to App Store Connect for testing; physical-device keyboard behaviour still needs validation.
#

iOS / Mac 1.0.0 · Build 53

App Store editions and purchase scope

  • Prepared iPhone/iPad and separate Mac App Store submissions with English, Traditional Chinese and Simplified Chinese store materials.
  • The Mac App Store edition copies completed dictation automatically; paste into another app with Command–V.
  • Clarified Free, Plus and Pro plans, separate iOS/Mac purchases and user-owned API costs. Submission is not approval or a public release.
#
Uploaded for testing

iOS 1.0.0 · Build 52

Microphone recovery and clearer guidance

  • Added a bounded retry when iOS reports that Siri is using the microphone, with actionable guidance when the microphone remains unavailable.
  • Cancellation and background checks prevent an abandoned activation from restarting. Applies to dictation, keyboard activation, meetings and live translation.
#
Uploaded for testing

iOS 1.0.0 · Builds 49–50

iPad layouts and safer writing retries

  • Added iPad-adaptable navigation, wider recording controls and side-by-side captions.
  • Compare multiple writing styles from the original transcript while retaining the last successful result. Optimized language-model processing, numeric accuracy and Cantonese text cleanup.
  • Improved recovery when writing processing needs a retry, while retaining the selected processing mode and existing download choices.
#
Uploaded for testing

iOS 1.0.0 · Builds 45–48

Translation captions and saved recordings

  • Added foreground, phrase-based translation previews with bilingual captions. This is not word-by-word simultaneous translation.
  • Added saved translation audio and captions, playback, sharing, copying and forwarding a recording to Notes.
  • Centralized API settings and simplified Notes/Translate screens. Language support depends on the selected route; the initial local translation preview did not expose Cantonese.
#
Uploaded for testing

iOS 1.0.0 · Builds 40–45

Meeting processing and local AI setup

  • Improved long-recording processing and reusable meeting checkpoints, and expanded Local AI controls for speech, writing and optional speaker models.
  • Added optional local speaker grouping and clearer model/download choices. Speaker labels are estimated groups, not verified identities.
  • Meeting access and device/model requirements apply; the current per-recording limit is 60 minutes.
#
Uploaded for testing

iOS 1.0.0 · Builds 34–37

Meeting notes, recovery and custom styles

  • Added a dedicated meeting-notes workflow, audio import, transcription and note organization, with saved progress and separate resume actions.
  • Added custom writing styles, industries and prompts for eligible plans, along with usage statistics and user-configured API options.
  • Added Lock Screen/Dynamic Island activity and bounded background work. Background operation remains subject to iOS scheduling and is not a promise of uninterrupted recording.
#
Uploaded for testing

iOS 1.0.0 · Builds 23–33

Preserve originals and improve result recovery

  • Improved writing-result validation, rejected-result feedback and recovery, while keeping the original transcript available for comparison.
  • Refined region/language onboarding and follow-system Chinese interface selection. Uploaded builds were for testing; store approval is a separate step.
#
Beta delivered

Mac 0.9.0 · direct beta

Confirmed wording rules and problem reports

  • Added user-confirmed contextual wording rules, a local rule manager and test preview.
  • Added opt-in sharing of local problem reports and refined Formal/Email writing instructions. Original transcripts remain preserved.
  • The signed and Apple-notarized direct-download Mac beta was published on the website. This is separate from the Mac App Store edition.
#
Package prepared

Mac 0.8.0 · direct beta

Sentence corrections and multilingual setup

  • Added sentence correction from results or History while preserving the original, plus regional speech preferences and up to three frequent languages.
  • Added Everyday/Precision package management and English, Traditional Chinese and Simplified Chinese interface preferences. Signed and notarized package prepared; later superseded by the published 0.9 beta.
#
Beta delivered

iOS 0.5.0–0.7.0

Offline packages and resilient downloads

  • Added region and frequent-language choices, punctuation-only mode, mixed-language safeguards and interface-language selection.
  • Added two offline packages, download size/speed/progress, retry/resume and background downloads, plus switching between retained downloaded packages.
  • Improved keyboard microphone activation, readiness countdown and return guidance. Delivered to the existing internal TestFlight beta group.
#
Beta delivered

VoiceType 0.2–0.3 → EnterSpeak 0.4

The first voice keyboard and EnterSpeak identity

  • Established app dictation, local history and offline-model workflows, then repaired recording and added keyboard voice input.
  • Renamed VoiceType to EnterSpeak and introduced a compact keyboard with microphone, status and session controls. Existing app data and model paths were retained.
#

Website updates

Published

Website

A multilingual changelog

  • Added a public development and release history in English, Traditional Chinese and Simplified Chinese, with clear delivery-status labels and navigation links.
#
Published

Website

Android privacy disclosures

  • Expanded the privacy policy to explain Android recordings, optional audio/text API sharing, encrypted API-key storage, keyboard context checks and backup/export behaviour.
#
Published

Website

Three languages, use cases and support

  • Published English, Traditional Chinese and Simplified Chinese pages, ten use-case guides, interactive preset writing examples and accessible motion controls.
  • Added shared support chat and clearer one-time purchase terms covering platform scope, API costs, restoration and existing-subscription cancellation.
#