Poor performance

[dan-burt]

Description

Infrequently use the app, though it is always running. Updated to v1.5.0 last week.

When I tried using it today, it started by saying "Loading model" for WhisperKit. never seemed to completed.

I had to delete this Extension, quit the app, restart it, then add WhisperKit back in. This still took many mins to fully load. This worked on the first occasion, as was very spritely in performance.

10 mins later, and retried a few times, back to no output. It reports "No model loaded", but the Settings page says otherwise. Also tried connecting to Airpods, same outcome.

Steps to Reproduce

Infrequently use the app, though it is always running. Updated to v1.5.0 last week.

When I tried using it today, it started by saying "Loading model" for WhisperKit. never seemed to completed.

I had to delete this Extension, quit the app, restart it, then add WhisperKit back in. This still took many mins to fully load. This worked on the first occasion, as was very spritely in performance.

10 mins later, and retried a few times, back to no output. It reports "No model loaded", but the Settings page says otherwise. Also tried connecting to Airpods, same outcome.

Expected Behavior

Working transcripts.

macOS Version

15.7.7

TypeWhisper Version

1.5.0

Diagnostics / Logs

{
  "api" : {
    "enabled" : false,
    "loopbackOnly" : true,
    "port" : 8978,
    "remoteAccessAllowed" : false,
    "running" : false
  },
  "app" : {
    "build" : "926",
    "bundleIdentifier" : "com.typewhisper.mac",
    "bundlePathKind" : "systemApplications",
    "isDevelopment" : false,
    "launchTimestamp" : "2026-07-04T13:35:31Z",
    "uptimeSeconds" : 1232.8790650367737,
    "version" : "1.5.0"
  },
  "audio" : {
    "audioDuckingEnabled" : false,
    "audioDuckingLevel" : 0.2,
    "defaultOutput" : {
      "deviceID" : 113,
      "name" : "MacBook Air Speakers",
      "transportType" : "builtIn",
      "uid" : "BuiltInSpeakerDevice",
      "volume" : 0.43749997
    },
    "inputDiagnostics" : {
      "defaultInputDeviceID" : 120,
      "devices" : [
        {
          "compatibility" : "unknown",
          "deviceID" : 103,
          "inputChannels" : 2,
          "inputOnlyCaptureFormat" : "48000.0 Hz\/2 ch\/pcmFloat32\/interleaved=false",
          "isAggregate" : false,
          "isAggregateOrVirtual" : true,
          "isDefaultInput" : false,
          "isSelected" : false,
          "isVirtual" : true,
          "listedByTypeWhisper" : true,
          "name" : "Microsoft Teams Audio",
          "nominalSampleRate" : 48000,
          "outputChannels" : 2,
          "transportType" : 1986622068,
          "transportTypeFourCC" : "virt",
          "transportTypeName" : "virtual",
          "uid" : "MSLoopbackDriverDevice_UID"
        },
        {
          "compatibility" : "unknown",
          "deviceID" : 120,
          "inputChannels" : 1,
          "inputOnlyCaptureFormat" : "48000.0 Hz\/1 ch\/pcmFloat32\/interleaved=false",
          "isAggregate" : false,
          "isAggregateOrVirtual" : false,
          "isDefaultInput" : true,
          "isSelected" : false,
          "isVirtual" : false,
          "listedByTypeWhisper" : true,
          "name" : "MacBook Air Microphone",
          "nominalSampleRate" : 48000,
          "outputChannels" : 0,
          "transportType" : 1651274862,
          "transportTypeFourCC" : "bltn",
          "transportTypeName" : "builtIn",
          "uid" : "BuiltInMicrophoneDevice"
        },
        {
          "compatibility" : "unknown",
          "deviceID" : 125,
          "inputChannels" : 1,
          "inputOnlyCaptureFormat" : "48000.0 Hz\/1 ch\/pcmFloat32\/interleaved=false",
          "isAggregate" : false,
          "isAggregateOrVirtual" : false,
          "isDefaultInput" : false,
          "isSelected" : false,
          "isVirtual" : false,
          "listedByTypeWhisper" : true,
          "name" : "iPhone (2) Microphone",
          "nominalSampleRate" : 48000,
          "outputChannels" : 0,
          "transportType" : 1667463012,
          "transportTypeFourCC" : "ccwd",
          "transportTypeName" : "unknown(1667463012)",
          "uid" : "5A9B7BC9-8314-477A-BD69-EB9E00000003"
        },
        {
          "compatibility" : "notListed",
          "deviceID" : 161,
          "exclusionReason" : "nameMatchedCADefault",
          "inputChannels" : 1,
          "inputOnlyCaptureFormat" : "48000.0 Hz\/1 ch\/pcmFloat32\/interleaved=false",
          "isAggregate" : true,
          "isAggregateOrVirtual" : true,
          "isDefaultInput" : false,
          "isSelected" : true,
          "isVirtual" : false,
          "listedByTypeWhisper" : false,
          "name" : "CADefaultDeviceAggregate-58840-0",
          "nominalSampleRate" : 48000,
          "outputChannels" : 2,
          "transportType" : 1735554416,
          "transportTypeFourCC" : "grup",
          "transportTypeName" : "aggregate",
          "uid" : "CADefaultDeviceAggregate-58840-0"
        }
      ],
      "previewActive" : false,
      "previewAudioLevel" : 0,
      "previewRawLevel" : 0,
      "selectedInputUsesBluetoothTransport" : false
    },
    "mediaPauseEnabled" : false
  },
  "counts" : {
    "dictionaryCorrections" : 0,
    "dictionaryTerms" : 0,
    "enabledProfiles" : 0,
    "enabledSnippets" : 0,
    "errorEntries" : 10,
    "historyRecords" : 119,
    "profiles" : 0,
    "snippets" : 0
  },
  "errors" : [
    {
      "category" : "transcription",
      "message" : "No model loaded. Please download and select a model first.",
      "timestamp" : "2026-07-04T13:37:12Z"
    },
    {
      "category" : "transcription",
      "message" : "No model loaded. Please download and select a model first.",
      "timestamp" : "2026-07-04T13:23:43Z"
    },
    {
      "category" : "transcription",
      "message" : "No model loaded. Please download and select a model first.",
      "timestamp" : "2026-07-04T13:19:47Z"
    },
    {
      "category" : "transcription",
      "message" : "No model loaded. Please download and select a model first.",
      "timestamp" : "2026-07-04T13:17:47Z"
    },
    {
      "category" : "transcription",
      "message" : "No model loaded. Please download and select a model first.",
      "timestamp" : "2026-07-04T13:16:19Z"
    },
    {
      "category" : "transcription",
      "message" : "No model loaded. Please download and select a model first.",
      "timestamp" : "2026-07-04T11:29:14Z"
    },
    {
      "category" : "transcription",
      "message" : "No model loaded. Please download and select a model first.",
      "timestamp" : "2026-07-04T10:43:55Z"
    },
    {
      "category" : "transcription",
      "message" : "No model loaded. Please download and select a model first.",
      "timestamp" : "2026-07-04T10:42:15Z"
    },
    {
      "category" : "transcription",
      "message" : "No model loaded. Please download and select a model first.",
      "timestamp" : "2026-04-23T08:57:47Z"
    },
    {
      "category" : "transcription",
      "message" : "No model loaded. Please download and select a model first.",
      "timestamp" : "2026-04-23T08:54:23Z"
    }
  ],
  "exportedAt" : "2026-07-04T13:56:04Z",
  "memoryExtraction" : {
    "inFlight" : false,
    "skippedWhileBusy" : 0,
    "totalCompleted" : 0,
    "totalFailed" : 0,
    "totalStarted" : 0
  },
  "model" : {
    "allowsTranscriptPreviewFallback" : true,
    "isModelReady" : false,
    "selectedModelId" : "openai_whisper-large-v3_turbo",
    "selectedProviderId" : "whisper",
    "supportsLiveTranscriptionSession" : false,
    "supportsStreaming" : true,
    "supportsTranslation" : true
  },
  "modelAutoUnload" : {
    "entries" : [

    ],
    "policyName" : "afterSeconds",
    "policySeconds" : 600
  },
  "permissions" : {
    "accessibilityGranted" : true,
    "microphoneGranted" : true
  },
  "plugins" : [
    {
      "allowsTranscriptPreviewFallback" : true,
      "bundled" : false,
      "enabled" : true,
      "id" : "com.typewhisper.parakeet",
      "isConfigured" : false,
      "name" : "Parakeet",
      "providerId" : "parakeet",
      "runtimeLoaded" : true,
      "source" : {
        "bundleIdentifier" : "com.typewhisper.parakeet",
        "bundleShortVersion" : "1.2.10",
        "bundleVersion" : "1",
        "executableSHA256" : "ddc894fd8953a9199b33e1a718aa69365beaaf19464aa485723d09bb2e85d130",
        "executableSizeBytes" : 11705840,
        "pathHint" : "~\/Library\/Application Support\/TypeWhisper\/Plugins\/ParakeetPlugin.bundle",
        "sourceKind" : "managedExternal"
      },
      "storedLoadedModelId" : "parakeet-tdt-0.6b-v2",
      "storedSelectedVersion" : "v2",
      "supportsLiveTranscriptionSession" : false,
      "supportsStreaming" : false,
      "version" : "1.2.10"
    },
    {
      "allowsTranscriptPreviewFallback" : true,
      "bundled" : false,
      "currentSettingsActivity" : {
        "isError" : false,
        "message" : "Loading model"
      },
      "enabled" : true,
      "id" : "com.typewhisper.whisperkit",
      "isConfigured" : false,
      "name" : "WhisperKit",
      "providerId" : "whisper",
      "runtimeLoaded" : true,
      "selectedModelId" : "openai_whisper-large-v3_turbo",
      "source" : {
        "bundleIdentifier" : "com.typewhisper.whisperkit",
        "bundleShortVersion" : "1.0.24",
        "bundleVersion" : "1",
        "executableSHA256" : "678f49ab67a02ea8587230d2c3eaef4710b48e44348a5096eb4915af27942293",
        "executableSizeBytes" : 6492176,
        "pathHint" : "~\/Library\/Application Support\/TypeWhisper\/Plugins\/WhisperKitPlugin.bundle",
        "sourceKind" : "managedExternal"
      },
      "storedLoadedModelId" : "openai_whisper-large-v3_turbo",
      "storedSelectedModelId" : "openai_whisper-large-v3_turbo",
      "supportsLiveTranscriptionSession" : false,
      "supportsStreaming" : true,
      "version" : "1.0.24"
    }
  ],
  "processMemory" : {
    "physicalFootprintBytes" : 316350336,
    "residentSizeBytes" : 879886336
  },
  "schemaVersion" : 7,
  "settings" : {
    "appFormattingEnabled" : true,
    "bundledReleaseChannel" : "stable",
    "dockIconBehaviorWhenMenuBarHidden" : "keepVisible",
    "externalStreamingDisplayCount" : 0,
    "historyRetentionDays" : 0,
    "indicatorStyle" : "overlay",
    "indicatorSupportsTranscriptPreview" : true,
    "indicatorTranscriptPreviewAvailable" : true,
    "indicatorTranscriptPreviewEnabled" : true,
    "indicatorTranscriptPreviewFontSizeOffset" : 0,
    "memoryCaptureScope" : "allDictations",
    "memoryEnabled" : false,
    "modelAutoUnloadPolicy" : "afterSeconds",
    "modelAutoUnloadSeconds" : 600,
    "notchIndicatorDisplay" : "activeScreen",
    "notchIndicatorVisibility" : "duringActivity",
    "overlayPosition" : "top",
    "saveAudioWithHistory" : false,
    "selectedLanguage" : "en",
    "selectedUpdateChannel" : "stable",
    "setupWizardCompleted" : true,
    "showMenuBarIcon" : true,
    "soundFeedbackEnabled" : true,
    "spokenFeedbackEnabled" : false,
    "translationEnabled" : false,
    "watchFolderAutoStart" : false
  },
  "skippedExternalBundles" : [

  ],
  "system" : {
    "cpuArchitecture" : "arm64",
    "localeIdentifier" : "en_GB",
    "macOSVersion" : "Version 15.7.7 (Build 24G720)",
    "timeZoneIdentifier" : "Europe\/London"
  },
  "workflows" : {
    "defaultLLMProviderId" : "appleIntelligence",
    "enabledCount" : 0,
    "enabledWorkflows" : [

    ],
    "totalCount" : 0
  }
}

[SeoFood]

Thanks for the diagnostics. I found that the public plugin feed was still serving WhisperKitPlugin 1.0.24, while TypeWhisper 1.5.0 expects the newer WhisperKit lifecycle fixes from 1.1.0.

I released WhisperKitPlugin 1.1.0 and verified that the public plugin registry now points to it:

• Release: https://github.com/TypeWhisper/typewhisper-mac/releases/tag/plugin-whisperkit-v1.1.0
• Registry entry: com.typewhisper.whisperkit -> 1.1.0
• Minimum host version: 1.5.0

Please open TypeWhisper > Integrations and update WhisperKit to 1.1.0. If the update does not appear immediately, remove and re-add the WhisperKit integration once so the app refreshes the installed bundle.

If it still gets stuck at “Loading model” or reports “No model loaded” after updating to 1.1.0, please attach a fresh diagnostics export and we will continue with the host-side restore timeout fix.

[dan-burt]

Thanks @SeoFood

Found the available WhisperKit update and applied v1.1.0.

No improvement. Tried immediately, and after a system reboot. Same outcome - long "Transcribing..." in the App's pill, with a final "No model selected behaviour".

Last time when it did fire into life, the transcription was quick. so it's obvious when it occurs. now trying few words as a simple test.

Last diags attached.

typewhisper-diagnostics-2026-07-04T16-22-22Z.json

[SeoFood]

Thanks for the follow-up diagnostics. I found that WhisperKit 1.1.0 fixed the public plugin feed mismatch, but not this restore path.

The fresh diagnostics show the selected/stored WhisperKit model is still openai_whisper-large-v3_turbo, while the plugin is still busy loading/optimizing and the host gives up too early, which then surfaces as the generic "No model loaded" error.

I opened #830 with the host-side restore wait fix and a WhisperKit guard so repeated restores for the same model do not restart the in-flight load. Once that PR is merged and released, this should go out as a TypeWhisper 1.5.x host fix plus WhisperKitPlugin 1.1.1.

[SeoFood]

Update: #830 has been merged and released.

• TypeWhisper 1.5.1: https://github.com/TypeWhisper/typewhisper-mac/releases/tag/v1.5.1
• WhisperKitPlugin 1.1.1: https://github.com/TypeWhisper/typewhisper-mac/releases/tag/plugin-whisperkit-v1.1.1
• The stable Sparkle appcast now points to TypeWhisper 1.5.1, and the public plugin feed now points to WhisperKitPlugin 1.1.1.

Please update TypeWhisper to 1.5.1, then open TypeWhisper > Integrations and update WhisperKit to 1.1.1. After both updates, try the same short dictation test again. If it still gets stuck, please attach a fresh diagnostics export from after both versions are installed.

[dan-burt]

updates applied, but the same issue. tried rebooting as well, same occurs after.

typewhisper-diagnostics-2026-07-05T10-31-41Z.json

[SeoFood]

Thanks, the new diagnostics were very useful. They confirm TypeWhisper 1.5.1 and WhisperKitPlugin 1.1.1 are installed, but show a different remaining state: the selected WhisperKit model is still openai_whisper-large-v3_turbo, while the stored loaded-model marker is missing.

I opened and merged #843 for that case. The fix makes WhisperKit restore from the selected model when the loaded-model marker is missing, instead of immediately falling through to the generic "No model loaded" error.

WhisperKitPlugin 1.1.2 is now released and the public plugin feed points to it:

• Release: https://github.com/TypeWhisper/typewhisper-mac/releases/tag/plugin-whisperkit-v1.1.2
• Registry entry: com.typewhisper.whisperkit -> 1.1.2
• Minimum host version: 1.5.0

Please open TypeWhisper > Integrations and update WhisperKit to 1.1.2. You do not need a new app update beyond TypeWhisper 1.5.1 for this specific fix. If it still gets stuck after installing WhisperKit 1.1.2, please attach one more fresh diagnostics export.

[SeoFood]

Hi @dan-burt, just checking in: did updating WhisperKit to 1.1.2 resolve the "No model loaded" / stuck transcription issue?

If it is fixed, I can close this issue. If it is still happening, please attach one more fresh diagnostics export after confirming TypeWhisper is on 1.5.1 and WhisperKit is on 1.1.2.

[dan-burt]

Actually the experience I had seemed to relate to the WhisperKit model I was using. I switched to the Parakeet model, which is downloading and transcripting properly, though it doesn't pick up punctuation as well.

Whisperkit just has the same problem. Saw it trying to "optimize" the model after the engine update, but same end-result.