joewongjc/type4me
60.4
Adequate · 15 November 2026
98.8k
lines of production code
Swift
primary language
5
measurements over time
What this system is
Features
Added LLM benchmarking scripts for voice polish and translation
Added Python scripts to the \scripts/\ directory to benchmark Large Language Model performance on voice polish and translation tasks. The suite includes \benchmark\_llm.py\ (comparing local Qwen3-4B vs. cloud Doubao), \benchmark\_4b\_vs\_9b.py\ (comparing Qwen3-4B vs. Qwen3.5-9B), \benchmark\_9b\_vs\_doubao.py\ (comparing Qwen3.5-9B vs. Doubao on longer samples), \benchmark\_gemini\_retest.py\ (testing Gemini prompt variations), and \benchmark\_llm\_quality.py\ (multi-model comparison via Ark and OpenRouter). These tools evaluate latency, output quality, and prompt effectiveness using samples from the app's recognition history.
scripts · high confidence
Expanded ASR provider support and unified configuration UI
The ASR provider configuration layer has been significantly expanded to support a wider range of speech-to-text services. New providers added include Apple Speech, AssemblyAI, Baidu AI Cloud, Alibaba Cloud Bailian, Cartesia, Deepgram, ElevenLabs, Gemini, Grok (xAI), Meta Muse, Xiaomi MiMo, Sherpa (local SenseVoice), Soniox, and StepFun (both real-time and batch). Existing providers have also been updated: OpenAI now supports model selection (GPT-4o Transcribe, GPT-4o Mini, Whisper) and custom base URLs; Volcano supports both new API Key and legacy App ID authentication with auto-detection for model resources; and AWS, Alibaba Cloud, Azure, Custom, Google, iFLYTEK, and Tencent providers now have localized UI labels for their credential fields. Also in this area: ASR providers now support selectable streaming or batch recognition modes.
Type4Me/ASR/Providers · high confidence
Expanded hotkey support for media keys, mouse buttons, and side-specific modifiers
The hotkey system now supports binding media keys (play, pause, volume, etc.) and mouse buttons as triggers, in addition to standard keyboard keys. It also improves modifier key handling by distinguishing between left and right side modifiers (e.g., left vs. right Command or Shift), ensuring more precise hotkey matching and preventing conflicts with system actions like Apple Music auto-launch. This allows users to create more flexible and context-aware hotkey bindings for voice input modes.
Type4Me/Input · high confidence
Expanded usage analytics and Ask Anything conversation history
The database layer now supports detailed LLM usage analytics and a new Ask Anything conversation history. For LLM usage, the system tracks token counts, costs, and request status in a dedicated \llm\_usage\_history\ table, enabling users to view usage breakdowns by model, feature, and date, as well as recalculate historical costs. Additionally, a new \AskAnythingStore\ manages conversation sessions and turns, allowing users to review and interact with their Ask Anything history.
Type4Me/Database · high confidence
New compact recording indicator with live transcript and correction learning
The floating bar now supports a compact recording style that displays a scrolling audio waveform history and a single-line live transcript that follows the text tail when it overflows. A new correction learning panel appears above the bar to let users accept or ignore transcription candidates with a 12-second auto-ignore timer. Additionally, a manual input panel allows users to type text directly with mode selection, and the floating bar now respects a light/dark recording theme for legibility across different host pages.
Type4Me/UI/FloatingBar · high confidence
New permission guide window and recording control coordination
The app now includes a dedicated 'Type4Me Permissions' window that appears during the setup flow, managed by a new PermissionGuideModel. Additionally, a RecordingControlCoordinator has been introduced to unify how recording actions (start, stop, follow-up) are handled, distinguishing between standard recording actions and those specific to the SelectionAsk feature. This change also adds environment objects for the app updater, menu bar control center, and navigation model to the main app structure, and introduces start gates to manage recording initiation logic more robustly.
Type4Me · high confidence
New streaming ASR providers: Apple, AssemblyAI, Baidu, Bailian, Cartesia, Deepgram, ElevenLabs, and Gemini
Type4Me adds support for eight new real-time speech-to-text providers, allowing users to transcribe audio via Apple Speech, AssemblyAI, Baidu, Alibaba Cloud Bailian, Cartesia Ink-2, Deepgram, ElevenLabs Scribe, and Google Gemini Transcribe. Each provider is implemented as a dedicated client (e.g., AppleASRClient, AssemblyAIASRClient) that manages its own WebSocket connection, handles authentication, and streams recognition events to the app. Also in this area: ASR providers now support selectable streaming or batch recognition modes.
Type4Me/ASR · high confidence
New translation language model and core service infrastructure
The app introduces a new TranslationLanguage model defining 18 supported languages with localized display names and stable prompt identifiers, alongside foundational services including AppBuildInfo for version display, AppDataLocation for separating user profile data from runtime state, AppUpdater for in-app updates, AskAnythingContextBuilder and Coordinator for managing conversation history and context truncation, AudioKeepAliveManager for maintaining microphone connections, BatchCorrectionInference for learning from user edits, ChineseWordSegmenter for text processing, CorrectionAffinityAnalyzer for evaluating correction confidence, CorrectionLearning for tracking and applying text corrections, and DataBackupManager for rotating local snapshots of user data. Also in this area: ASR providers now support selectable streaming or batch recognition modes.
Type4Me/Services · high confidence
Redesigned Settings UI with live appearance preview and conversation history management
The Settings interface has been completely redesigned to improve usability and visual feedback. A new Appearance tab now features a live preview stage that renders the recording indicator and text output formatting in real-time as you adjust theme, style, and formatting options. The app also introduces an 'Ask Anything' page, allowing users to search, browse, and continue past voice-driven conversation sessions. Additionally, the settings now include a provenance explanation for history records, clarifying why a record's output differs from its raw recognition, and a new Debug tab provides build diagnostics and log access for troubleshooting. Also in this area: Fix voice input in Type4Me settings and remove dark flash on hover.
Type4Me/UI/Settings · high confidence
User-selectable recording start sounds and improved audio feedback reliability
Users can now choose from eight distinct start-sound styles (including Chime, Pluck, Submerge, Pong, and Water Drops) via the \tf\_startSound\ setting, replacing the previous single synthesized tone. The \SoundFeedback\ module has been refactored to use pre-warmed \AVAudioPlayer\ instances and cached PCM buffers, which eliminates frame-drop issues and ensures consistent audio playback timing for start, stop, and error events.
Type4Me/Session · high confidence
Removals
Removal of AECTest verification tool and SpeexDSP dependency
The AECTest verification tool, which recorded microphone and system audio to validate acoustic echo cancellation, has been removed along with the embedded CSpeexDSP library sources and headers. This eliminates the ability to run the standalone AEC verification workflow and removes the bundled SpeexDSP dependency from this location.
AECTest · high confidence
Removal of AudioCaptureEngine
The \AudioCaptureEngine\ class, which previously handled audio capture using \AVFoundation\ (including warm-up, start/stop logic, and sample buffer conversion), has been deleted from the codebase.
Type4Me/Audio · high confidence
Removal of Claude and Doubao LLM provider implementations
The application no longer supports the Claude (Anthropic) and Doubao (ByteDance ARK) LLM providers. The specific client implementations (\ClaudeChatClient\, \DoubaoChatClient\), the shared \LLMClient\ protocol, the \LLMProvider\ enum, and the \LLMProviderConfig\ protocol have been removed from the codebase. Users will no longer see these options in the model selection list and cannot configure credentials for them.
Type4Me/LLM · high confidence
Removed AEC demo tool
The AEC demo tool (aec-demo.swift) has been removed from the tools directory. This tool previously allowed users to record audio samples with and without acoustic echo cancellation to compare performance.
tools · high confidence
Behavioural changes
4 commits (0 fixes) modifying Type4Me/Resources
A change to existing behaviour in Type4Me/Resources — 4 commits, 1 file: “Initial commit: Type4Me - macOS voice input tool”; “feat(brand): 新增「4!」小尺寸 Logo,并用于菜单栏与应用图标 (\#351)”; “Update AppIcon.icns”; and 1 more (refs \#351).
Type4Me/Resources · medium confidence · unverified
Granular clipboard retention and safer text injection
Users now have explicit control over whether dictated text remains on the system clipboard after completion or cancellation via the new Clipboard Output Policy settings (Always Copy, On Cancel: Processed, On Cancel: Raw Transcript, Never Copy). The injection engine has been refactored to use a snapshot-and-restore model that preserves the user's previous clipboard content, and internal clipboard writes are now marked as transient to prevent cluttering third-party clipboard history apps. Additionally, text injection now supports tracked replacements with verification to ensure text is correctly applied to the focused field, improving reliability in complex applications.
Type4Me/Injection · high confidence
Improved dark-mode contrast and new correction feature
The dark-mode recording indicator text is now fully opaque white instead of semi-transparent, fixing low contrast on bright backgrounds. A new correction feature allows users to select a misrecognized word, hold to speak the correct text, and have the app replace it (and similar occurrences) using an LLM. Additionally, the Volcano (Doubao) ASR provider and Doubao LLM provider have been renamed to Volcengine and Volcengine Ark respectively, and the configuration guide link for Volcano ASR now points to a GitHub issue. Also in this area: Modifier keys can now use double-tap to trigger recording; Expanded speech recognition providers and improved macOS permission UX.
Sources · high confidence
Keyboard interaction overhaul: redo, native mode switcher, and haptics
The iOS keyboard now supports redoing an undone insertion via a new side key, replacing the previous undo-only flow. The mode picker has been replaced with a native, trackless segmented control that matches the iOS Liquid Glass aesthetic and allows dragging between modes. Key presses now provide tactile and audio feedback (haptics and system sounds) configurable in settings, and the punctuation key supports flicking for a full stop and holding to select other marks. Additionally, the keyboard's display name and URL scheme are now configurable via Signing.local.xcconfig for custom builds. Also in this area: Add optional keyboard clipboard suggestion; iOS 1.1.0: New correction feature, text output settings, and UI overhaul; Intelli Sense for iPhone, customizable app icons, and configurable build identity; New app icon selection, keyboard feedback controls, and improved session navigation; App Store submission prep and TestFlight public testing.
iOS · high confidence
Redesigned floating bar with new visual styles and layout options
The floating recording bar now supports two layout modes—Regular and Compact—allowing users to choose a standard capsule or a smaller indicator. A new set of visual styles (such as Siri Ripple, Frost Fluid, and Liquid Chrome) replaces the previous defaults, with a power-saving static option available. The UI also introduces configurable settings to show or hide tooltips, the cancel button, and the finish button, while ensuring the bar remains legible over dark host pages via improved contrast floors for the native frosted glass effect.
Type4Me/UI · high confidence
Release v3.1.0 and updated README header visuals
The build scripts (build-dmg.sh, package-app.sh) now default to version 3.1.0, preparing the application for this release. Additionally, the README header generation script (scripts/readme/gen\_header.py) has been updated to render the new 'Liquid Glass' recording bar design, including specific light and dark theme colors, a handwritten wordmark, and an animated glyph-streaming effect that mirrors the app's FloatingBarView behavior. Also in this area: Automated bilingual README hero images.
scripts/readme · high confidence
Setup wizard no longer shows demo steps for Quick and Custom modes
The setup wizard has been streamlined by removing the interactive demo steps that previously illustrated Quick Mode and Custom Mode. Users will no longer see the animated text-transcription previews or the simulated FloatingBar demonstration during the initial configuration flow, resulting in a more direct path to permissions and provider setup.
Type4Me/UI/Setup · high confidence
Test coverage
Added tests for correction core, mobile config export, recording glass tuning, and transfer payload handling; Added vocabulary ASR accuracy test script; Removed obsolete test suite files.
Dependencies
Introduces conditional build targets and evaluation harness for local ASR and IntelliSense
The main Package.swift now supports conditional compilation for local speech recognition (SherpaOnnx), cloud subscription features, and Chinese text processing (CppJiebaBridge), allowing the application to be built with or without these capabilities. It also exposes new library targets (Type4MeUI, Type4MeIntelliSenseCore, Type4MeReviseCore) and lowers the minimum macOS deployment target to 14. Additionally, a new Evaluation/IntelliSenseEval package is added to test the IntelliSense core, and the qwen3-asr-server directory now includes a requirements.txt specifying dependencies like mlx, fastapi, and uvicorn for the local ASR server. Also in this area: Add Type4MeCorrectionCore module to package manifest; Project structure reorganization and test suite addition.
(dependencies) · high confidence
Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
How this codebase got here
Score
- CAI 41 → 60 (+19.3)
- Rubric changed (rubric-2026.10.5 → rubric-2026.11.4) — scores are not directly comparable.
Lenses
- Code Health 74 → 68 (-6.6)
- Architecture 92 → 93 (+1.1)
- Maturity 72 → 77 (+4.8)
- Readiness 43 → 56 (+12.9)
- Security 60 → 60 (+0.0)
- Performance 60 → 60 (-0.4)
Resolved (114)
- Duplicated block (10 lines × 2) (Sources/Type4Me/UI/FloatingBar/FloatingBarPlacementGeometry.swift)
- Duplicated block (10 lines × 2) (Sources/Type4MeReviseCore/ReviseInstructionAnalyzer.swift)
- Duplicated block (10 lines × 2) (iOS/App/UI/HomeView.swift)
- Duplicated block (11 lines × 2) (Sources/Type4Me/ASR/Providers/AzureASRConfig.swift)
- Duplicated block (11 lines × 2) (Sources/Type4Me/Audio/StartSoundRouteGate.swift)
- Duplicated block (11 lines × 2) (Sources/Type4Me/Protocol/MiMoASRProtocol.swift)
- Duplicated block (11 lines × 2) (Sources/Type4Me/Services/VolcBoostingTableManager.swift)
- Duplicated block (11 lines × 2) (Sources/Type4Me/UI/Settings/BrandIconView.swift)
- Duplicated block (11 lines × 3) (Sources/Type4MeReviseCore/ReviseScopeResolver.swift)
- Duplicated block (12 lines × 2) (Sources/Type4Me/ASR/ASRProvider.swift)
- Duplicated block (12 lines × 2) (Sources/Type4Me/ASR/Providers/GroqASRConfig.swift)
- Duplicated block (12 lines × 2) (Sources/Type4Me/LLM/ASRVariantGenerator.swift)
- Duplicated block (12 lines × 2) (Sources/Type4Me/Protocol/BatchASRTranscribers.swift)
- Duplicated block (12 lines × 2) (Sources/Type4Me/UI/Settings/History/ASRUsageAnalyticsView.swift)
- Duplicated block (12 lines × 2) (iOS/App/UI/ServiceSettingsView.swift)
- Duplicated block (12 lines × 3) (Sources/Type4Me/Database/HistoryStore.swift)
- Duplicated block (12 lines × 4) (Sources/Type4Me/Database/HistoryStore.swift)
- Duplicated block (12 lines × 4) (Sources/Type4Me/Database/HistoryStore.swift)
- Duplicated block (13 lines × 2) (Sources/Type4Me/Services/IntelliSenseSettings.swift)
- Duplicated block (13–31 lines × 2) (Sources/Type4Me/UI/AppState.swift)
- …and 94 more
New (198)
- ClassTooLong: VoiceSession (iOS/App/Pipeline/VoiceSession.swift)
- Duplicated block (10 lines × 2) (Sources/Type4Me/UI/FloatingBar/FloatingBarPlacementGeometry.swift)
- Duplicated block (10–11 lines × 3) (Sources/Type4MeReviseCore/ReviseScopeResolver.swift)
- Duplicated block (11 lines × 2) (Sources/Type4Me/Audio/StartSoundRouteGate.swift)
- Duplicated block (11 lines × 3) (Sources/Type4MeReviseCore/ReviseScopeResolver.swift)
- Duplicated block (12 lines × 2) (Sources/Type4Me/UI/Settings/Home/HomeDashboardModels.swift)
- Duplicated block (13 lines × 2) (Sources/Type4Me/LLM/ASRVariantGenerator.swift)
- Duplicated block (13 lines × 2) (Sources/Type4Me/Services/IntelliSenseSettings.swift)
- Duplicated block (18 lines × 5) (Sources/Type4Me/ASR/AssemblyAIASRClient.swift)
- Duplicated block (19 lines × 2) (Sources/Type4Me/Services/KeychainService.swift)
- Duplicated block (41 lines × 2) (Sources/Type4Me/Session/RecognitionSession.swift)
- Duplicated block (5 lines × 2) (iOS/App/LLM/ChatClient.swift)
- Duplicated block (5 lines × 4) (Sources/Type4Me/ASR/AssemblyAIASRClient.swift)
- Duplicated block (6 lines × 2) (Sources/Type4Me/ASR/SpeechRecognizer.swift)
- Duplicated block (7 lines × 2) (Sources/Type4Me/ASR/AssemblyAIConnectionGate.swift)
- Duplicated block (7 lines × 3) (Sources/Type4Me/ASR/AzureMAIASRClient.swift)
- Duplicated block (8 lines × 2) (Sources/Type4Me/ASR/BaiduASRClient.swift)
- Duplicated block (9 lines × 2) (Sources/Type4Me/ASR/AssemblyAIASRClient.swift)
- Duplicated block (9 lines × 4) (Sources/Type4MeReviseCore/ReviseScopeResolver.swift)
- Duplicated block (9–10 lines × 2) (Sources/Type4Me/ASR/SonioxASRClient.swift)
- …and 178 more
Changes since last survey
- 8 commits — 6 feature/other, 2 fixes
By area
- (root) — 2 commits
- iOS/App — 2 commits
- (repo) — 1 commit
- Sources/Type4Me — 1 commit
- Tests/Type4MeTests — 1 commit
- iOS/Keyboard — 1 commit
Notable commits
- fix: fix(ios): keep the background mic with a video-call PiP, add keyboard diagnostics (#403)
- fix: fix(ios): 键盘快捷粘贴不再显示很久以前的复制 (#404)
- change: Merge pull request #401 from joewongjc/chore/ios-1.1.0-testflight-manifest
- change: chore: announce iOS 1.1.0 (3) on TestFlight
- change: feat(ios): explain when a call or another app holds the microphone (#405)
- change: feat(llm): 刷新文本处理模型列表,Claude 支持关闭思考 (#407)
- change: feat: upgrade ElevenLabs realtime ASR to Scribe v3 Live (#409)
- change: release: v3.2.0
Architecture
- Unchanged — 0 containers · 1 contexts · 0 edges
Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
Survey your own repository
joewongjc/type4me was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.
About this page
- The score is its most recent published measurement, taken on 15 November 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
- Measured at commit 28353340207c5288f9dcaf2e71928cc3487d49cf — the exact code this score is about.
- Scored under rubric-2026.11.4 — the same rubric and the same method as every other entry in this index.
- Measured by Canine Code Analyzer (cca.canine.dev) using cca measure-2026.11.65+preprod-3c33c1882dc1.