Caption Lane
Support
For help with Caption Lane for iPhone, email lopl@lopl.net. Include your iPhone model, iOS version, app version and what happened. Leave out private meeting content and payment details.
- Before your first recording, download the recording models over Wi-Fi. The download is about 506 MB.
- During a recording, tap CC for live English captions. Enable microphone and speech recognition access if iOS asks.
- Free recordings can be up to 10 minutes each. A one-time purchase removes that limit. Open Settings → Longer recordings → Restore Purchases to restore access with the same Apple Account.
Privacy policy
Caption Lane uses your iPhone’s microphone to record audio when you start a recording. Live English captions use on-device Apple Speech when you tap CC. The app transcribes saved recordings and separates speakers on your iPhone. Caption Lane does not upload recordings, live captions, or saved transcripts.
Recordings and saved transcripts stay in the app until you delete them. If you enable device backups, iOS may include the app’s stored meetings. To delete a meeting and its recording and transcript, swipe left on the meeting in the library and tap Delete. When you share a transcript, you choose the recipient or destination in the iOS share sheet.
The app contacts Hugging Face and its content delivery providers to download the recording models. These services receive ordinary connection information, such as an IP address. The download contains model files. Apple handles purchases and restores through StoreKit. Caption Lane receives purchase status so it can unlock longer recordings.
Caption Lane has no accounts, ads, or analytics. If you contact support, the developer uses the email address and details you provide to respond and keeps them as long as needed to handle your request. For privacy questions or to ask about deleting support correspondence, email lopl@lopl.net.
Speech software and model credits
- FluidAudio by Fluid Inference, under the Apache License, Version 2.0.
- NVIDIA Parakeet TDT 0.6B v3, converted to Core ML by Fluid Inference, under Creative Commons Attribution 4.0.
- Fluid Inference’s Core ML speaker diarization models, based on pyannote and WeSpeaker.
- Silero VAD, converted to Core ML by Fluid Inference.