Voice capture
Voice capture lets you record spoken notes directly in the Ultrathink iOS app. Your recordings are uploaded and automatically transcribed using AI, so you can search and browse your voice notes as text alongside all your other entries.
This feature is useful for capturing ideas when typing is not practical: while walking, driving, or whenever you need to get a thought down quickly.
Recording a voice note
Starting a recording
- Open the Ultrathink app on your iPhone
- Tap the microphone icon on the main screen
- Recording starts immediately; you will see a timer and a visual waveform indicating that audio is being captured
There is no separate "start" button. Tapping the microphone icon begins recording straight away, so you can start speaking as soon as you tap.
Stopping and saving
- Tap the stop button (the microphone icon changes to a stop icon while recording)
- The recording preview screen appears, showing the duration of your audio
- Enter a title for the voice note
- Optionally add notes to provide additional context
- Tap Save
The entry is created in your knowledge base with the audio file attached. It syncs to the cloud and begins AI processing automatically.
How transcription works
After you save a voice note, the audio file is uploaded to the Ultrathink cloud. The transcription process follows these steps:
- Upload: the audio file is sent to Firebase Storage (encrypted in transit)
- Transcription: the audio is processed by OpenAI Whisper, a speech-to-text AI model
- Processing: the transcribed text is treated like any other entry content; Ultrathink's AI generates a summary, extracts topics, and identifies relationships
- Completion: the entry updates with the full transcription, which you can view in the detail panel of the web app or the iOS app
Transcription typically completes within a few minutes, depending on the length of the recording and current server load.
Viewing transcriptions
Once transcription is complete, the text appears in the entry alongside the original audio:
- In the web app: open the entry and the transcribed text is displayed in the detail panel. The AI-generated summary is based on the transcription.
- In the iOS app: tap the entry to see the transcription and any AI-generated summary.
You can edit the transcription in the web app if corrections are needed. The AI summary updates based on the corrected text the next time the entry is processed.
Tips for good transcription quality
The quality of your transcription depends largely on the audio quality. Follow these guidelines for the best results:
Speaking and environment
| Tip | Why it helps |
|---|---|
| Speak clearly and at a natural pace | Rushed or mumbled speech produces more errors |
| Minimise background noise | Background conversations, music, and wind interfere with recognition |
| Hold the phone at a comfortable distance | Too close causes distortion; too far picks up ambient noise |
| Avoid covering the microphone | Cases or fingers over the mic muffle the audio |
Recording length
- Keep recordings under 10 minutes for the most reliable transcription results
- Longer recordings still work, but transcription accuracy may decrease and processing takes longer
- If you have a lot to say, consider breaking it into multiple shorter recordings, each focused on a single topic
Content structure
- State the topic at the beginning of your recording so the AI can generate an accurate title suggestion
- Pause briefly between ideas to help the transcription model identify sentence boundaries
- Spell out unusual names or technical terms if accuracy matters (for example, "Kubernetes, that is K-U-B-E-R-N-E-T-E-S")
- Avoid filler words where possible; while the model handles them well, fewer filler words mean a cleaner transcription
Microphone permissions
The first time you tap the microphone icon, iOS asks for permission to access the microphone. You must grant this permission for voice capture to work.
Granting permission
- When the permission dialog appears, tap Allow
- Recording begins immediately after you grant permission
If you previously denied permission
If you tapped "Don't Allow" by mistake, you can change this setting:
- Open the Settings app on your iPhone
- Scroll down and tap Ultrathink
- Tap Microphone
- Toggle the switch to ON
- Return to the Ultrathink app and tap the microphone icon again
Troubleshooting
Microphone icon does not respond
- Check that microphone permission is granted in iOS Settings (see the "Microphone permissions" section above)
- Make sure no other app is currently using the microphone (for example, a phone call or another recording app)
- Force-close the Ultrathink app and reopen it
Recording saves but transcription never appears
- Verify your internet connection; the audio must upload before transcription can begin
- Wait a few minutes; transcription can take time depending on server load
- If the transcription still does not appear after 15 minutes, try opening the entry in the web app and checking its status
Transcription contains many errors
- Review the tips in the "Tips for good transcription quality" section above
- Re-record in a quieter environment if background noise was the issue
- For technical or specialised vocabulary, consider adding a note with the correct terms so they appear in the entry alongside the transcription
- You can manually edit the transcription in the web app if needed
Audio quality is poor
- Make sure you are not covering the microphone with your fingers or a case
- Hold your iPhone approximately 15 to 25 centimetres from your mouth
- Avoid recording in windy conditions or near fans and air conditioning units
- If you are using Bluetooth headphones, try recording with the iPhone's built-in microphone instead for comparison