Krak-EN Voice
User Guide

Record a conversation, turn it into text, and find the details you need.

For Android, iPhone and iPad · Updated September 30, 2026

1. Get started

  1. Open the app and complete the welcome screens. Allow microphone access when you want to record.
  2. On iPhone or iPad: open Settings → Trial & Lifetime Unlock and choose Start 30-day Trial — Free. Complete Apple’s confirmation. An internet connection is needed to contact the App Store. On Android: start with the free edition; use Upgrade to Pro when you want the paid features.
  3. Check Settings → AI Models. Whisper provides speech transcription. Gemma provides AI summaries and chat; use Manage AI Models to check availability and download it on a supported device.
  4. Use Wi-Fi for model downloads, leave enough free device storage, and wait for setup to finish before starting AI work.
  5. Make a short test recording and play it back before relying on the app for an important conversation.

Recording and supported AI processing run on your device once the required models are installed. Gemma availability depends on your device’s memory and hardware; the model setup screen reports whether your device supports it. Recording and transcription do not require Gemma.

The Apple app requires iOS or iPadOS 17 or later. The Android edition requires Android 12 or later and supports selected Samsung phone hardware; a supported Android version alone does not guarantee compatibility. Android setup checks the device model and processor. PDF text import on Android additionally requires Android 15 or later. Follow the compatibility message in your edition before downloading models. A purchase does not change hardware requirements for local AI.

3. Record and check your microphone

  1. On Home, select the language spoken in the recording. Use Settings → Default Language to set your usual choice.
  2. Place the device near the speakers with its microphone unobstructed. Tap the microphone control to check the input device if you use an external microphone.
  3. Tap the large record button and allow microphone access if prompted.
  4. Speak normally. The timer should advance and the visualizer should respond to your voice.
  5. Use Pause and Resume to suspend and continue the same recording. Tap Stop to finish and save it.
  6. Open Files to find the saved recording. Depending on your Auto-transcribe setting, transcription starts automatically, is offered as a choice, or waits until you request it.

Input and sensitivity

Open Microphone Settings from Home to select an available input or System Default. The Sensitivity slider changes the visualizer’s response; it does not increase the recorded audio volume. Improve a quiet recording by moving the microphone closer and reducing background noise.

Both operating systems can interrupt audio when calls or other system events occur. On Android, allow requested phone-state access to help pause for calls, and keep the ongoing recording notification enabled. Android uses foreground services for ongoing work; battery settings and system restrictions can still interrupt it. After an interruption, check the saved recording and recording status before continuing. Only record material you have permission to capture.

4. Transcribe a recording

  1. Check that Whisper is ready in Settings → AI Models.
  2. Open a recording from Files and start transcription when offered. After recording or importing, you can also choose Transcribe in the prompt.
  3. Allow the job to finish. Longer recordings take longer; keep the app open during lengthy processing.
  4. Read the resulting transcript and compare uncertain passages with the original audio.

In Settings → Auto-transcribe, choose automatic transcription, Ask me, or Manual. For poor results, check the recording language and audio quality before using Re-transcribe. Names, numbers, accents, overlapping speakers, and background noise can cause errors.

Automatic speaker identification is not available in this release. Do not assume the transcript identifies who said each line.

5. Generate and review summaries

  1. Finish transcription and open the recording’s detail screen.
  2. Choose Generate AI Summary. If a model is required, open model management and complete Gemma setup first.
  3. Wait for processing to finish, then review the summary against the transcript and audio.
  4. Use the available refinement or regeneration controls if you need a different result.

Summaries can help you review key points, decisions, and follow-ups. AI can omit information or introduce mistakes. Check dates, names, quantities, and action items before relying on or sharing the result. Keep the app in the foreground for long AI tasks: iOS and iPadOS limit processing time in the background. Android uses foreground services for supported ongoing tasks, but keeping the app open is also useful when processing large files.

6. Ask about your recordings

  1. Return to Home and type in Ask about your recordings….
  2. For a question about one recording, tap the paperclip and select that recording. Its transcript becomes context for your question.
  3. Send a focused prompt such as “What decisions were made?” or “List the follow-ups mentioned in this recording.”
  4. Read the response and open any source recordings to check the details.
  5. Use the stop control to stop a response, or the conversation control to reopen the current chat. Use the save control when you want to save a chat summary.

Transcribe a recording before asking about what was said. Chat requires a ready Gemma model and any access required by your edition; on Apple devices this means an active trial or lifetime unlock. A pinned recording can be removed with its close control. Chats are temporary unless you save the work you want to keep.

7. Organize and import files

Open Files and use New Folder to create a folder. Open a recording to review it; use its rename and move controls to give it a useful name and destination. Descriptive titles make later searches easier.

  1. Tap Import Files in Files or inside a folder.
  2. Choose Import audio, then select an audio file in the system picker. If it is stored in iCloud Drive, Google Drive, or another file provider, download it first if needed.
  3. Choose a destination folder when prompted.
  4. Open the imported recording, check playback, and choose whether to transcribe it.

Import success depends on the file’s audio format and whether your device can read it. If a file fails, try an unprotected standard audio file and check that the download completed.

Import a PDF or Word document

  1. Tap Import Files → Import PDF or Word document.
  2. Select a PDF or modern Word .docx file smaller than 50 MB, then choose a folder if prompted.
  3. After import, open Extracted document text to check the text. Review any extraction warning.
  4. Choose Generate AI Summary when Gemma is ready and your edition permits AI processing.

PDF import supports up to 200 pages of readable text. Android PDF text extraction requires Android 15 or later; on an older supported Android phone, update Android if available or import a DOCX version instead. Scanned or image-only PDFs need OCR before import. If some pages have no text, the app reports how many were omitted. Password-protected PDFs need an unlocked copy.

Word import reads body text and tables. Images, comments, headers, footers, and footnotes are not included. Save older .doc files as .docx first. Documents exceeding 500,000 extracted characters must be split into smaller documents. DOCX files with more than 8 MB of uncompressed main-document XML also need to be reduced or split. Your source file is preserved; the app keeps its own imported copy.

8. Export and share

  1. Open a recording and tap Export.
  2. Select one or more available formats: PDF Document, Word Document, Audio File, or Transcript Only.
  3. Tap the export button and choose the destination in your device’s share sheet, such as Files, a file manager, Mail, Gmail, or another installed app.
  4. Open the exported file to confirm that it contains the content you intended to share.

PDF and Word exports need a transcript or summary. Transcript-only export needs a transcript. Audio export requires the original audio still to exist. Review Settings → Default Export Format and the export content controls for your preferred output. Branding controls let you customize reports when available with your access.

Sharing creates a separate copy. Deleting the recording later does not delete copies in Files, email, cloud storage, or a recipient’s device. Exports are also a practical way to keep important work outside the app.

9. Calendar reminders

Open Calendar and allow access if you want upcoming events from your device calendar to appear. The view shows upcoming meetings over the next seven days. Choose a reminder interval and tap the bell next to a meeting, or open an event and choose Remind Me to Record.

Allow notifications for reminders. On Android, allow alarms/reminders access if requested, and check battery restrictions if reminders arrive late. A reminder prompts you to open the app and start recording; it does not automatically record a meeting. If an event is missing, check its date and that its calendar is enabled on your device.

10. Retention, Trash, and backups

Review Settings → Default Retention Policy before making recordings you need to keep. A recording’s Audio Retention control provides these options:

Retention cleanup removes audio separately from its transcript and summary. Once audio is gone, it cannot be played or exported from the app. Apple trial expiry does not itself delete saved work. Android’s free edition also has a 2 GB audio storage cap: cleanup can remove least-recently-accessed audio when the cap is exceeded, separately from the selected retention policy. Export important audio before reaching the limit. The Android paid unlock removes this free-tier cap; audio retention settings still apply.

Restore or permanently delete

Deleted recordings go to Trash. Restore a recording there if you deleted it accidentally. Permanent deletion and emptying Trash remove items without an in-app undo. Trashed items may be cleaned up after 30 days; restore important items promptly.

Protect important work

The app does not provide its own cloud sync or remote recovery service. Export important recordings and documents before changing devices or removing the app. Device backups may include app data according to your Apple, Android, Google, or computer backup settings. Restoring a purchase restores access, not your recordings.

11. Trial, lifetime unlock, and offer codes

iPhone and iPad — Apple App Store

The Apple edition offers a free 30-day trial. After the trial, the lifetime unlock is a one-time US$9.99 purchase in the US; the app displays your local App Store price. There is no subscription or automatic charge when the trial ends.

Open Settings → Trial & Lifetime Unlock to check access. Starting a trial requires Apple’s confirmation; reinstalling the app does not create a new trial for the same Apple Account.

After trial expiry, new recording and AI processing require the unlock. You can still open, play, export, and delete saved work without buying it. A pending purchase grants access after Apple confirms it. If a completed purchase does not appear, connect to the internet and try Restore Purchases.

Android — Google Play

The Android free edition limits each recording to 20 minutes. At the limit, recording stops and saves; you can start a new recording or upgrade. Its 2 GB audio storage cap is explained in the retention section above. Use Upgrade to Pro for the one-time full unlock, including removal of those free recording/storage limits and available branding/export options. Check the Google Play purchase sheet for the current price in your currency. The Apple 30-day trial and Apple US$9.99 price do not describe Android’s offer.

Use Restore Purchase with the Google Account used for the original Google Play purchase. Complete pending payments in Google Play before expecting access. Apple offer codes cannot be redeemed in Google Play; any Android promotion must be valid for Google Play and follow the instructions provided with it.

Moving between platforms

Purchases belong to the store and account used to obtain them. The app does not automatically transfer Apple purchases to Android or Google Play purchases to Apple. Restoring an unlock does not transfer recordings, transcripts, or downloaded models.

12. Troubleshooting

The microphone graph does not move

Start recording, check that the timer advances, and speak near the selected microphone. Check microphone permission in your device settings, the input device, and the Sensitivity slider. Play back a short test. Install the latest app update if using an older build.

The keyboard hides Home controls

Tap Hide keyboard, tap outside the chat field, or switch tabs. On Android, use the system Back or keyboard-dismiss button.

Transcription or summaries do not start

Check the access required by your edition, model readiness, free storage, and whether another task is still processing. Gemma is needed for summaries and chat; Whisper is needed for transcription. Keep the app open and allow more time for long recordings. If a job fails, retain its source audio and retry after checking model setup.

A model download fails

Check the network and free device storage. Open Manage AI Models to read the error and retry the download. An unsupported-device message means the device cannot run that model; buying lifetime access will not remove that requirement.

Audio is missing but text remains

Check the recording’s retention policy. Audio may have been removed after transcription or after its retention period. Look for an export or backup; the app cannot reconstruct the original recording from text.

A purchase or code fails

Check your internet connection, the Apple Account or Google Account used to purchase, and any code expiry. Try Restore Purchases for access already purchased. If the Apple sheet reports Environment: Sandbox, you are using a test build. Development-build testers need a Sandbox Apple Account configured by the developer; contact support for test-account help. Do not send account passwords or verification codes.

13. Privacy and permissions

Your audio, transcripts, and supported AI work are processed locally. Model downloads, Apple App Store or Google Play transactions, opening website links, and sharing files can use the network. You choose what to export and where it goes.

The app uses a local encrypted database for structured records. Audio and exported files are separate, and the vault opens automatically in this version. Protect your device with a passcode and protect exported copies separately. Manage microphone, calendar, and notification permissions in your device settings.

Read the Privacy Policy for storage, backups, third-party services, support information, and deletion details.

14. Get help

Email info@nsflabs.org for app support, or use Settings → Report a Bug. For privacy questions, contact info@nsflabs.org.

Include your device model, Android/iOS/iPadOS version, app version/build from Settings, the exact error, and the steps that led to it. A screenshot can help; remove personal information before sending it. Do not include recordings, transcripts, passwords, or purchase codes unless you deliberately intend to share the relevant material.

You can print this guide or save it as a PDF using your browser’s print controls.