Ferry Translate ("Ferry", "the app") is developed by Zhang Wei for iPhone, iPad and Mac. In short: we do not collect, store or sell any of your personal data. Ferry has no account system, no analytics and no ads, and we run no server that receives your data.
Where your content goes
Text you enter, images or videos you take, paste or choose, speech or audio files you record or import, and the picture and sound recorded for subtitle images are handled in one of two ways, depending on your settings:
- On-device translation, text recognition, speech recognition, long-screenshot stitching and video subtitle extraction are performed locally by Apple's system frameworks and Ferry's on-device algorithms. The content does not leave your device.
- An AI service you configure. If you enter an API key for an AI service (for example OpenAI, DeepSeek, Mistral, OpenRouter or a self-hosted service), Ferry sends the content to be processed directly from your device to that service. It does not pass through us. How that service handles your content is governed by its own privacy policy and your agreement with it.
Translation, speech, voice recognition and image recognition can each use a different service. Anything you haven't configured uses the system's built-in capabilities.
When you share a web link to Ferry, Ferry fetches that page to read its title and text for translation. Links to posts on X (Twitter) are read through X's public embed endpoint with a Do Not Track flag. When you share a link to an image, Ferry downloads that image to recognize its text.
Data stored on your device
- Settings and cached model lists, such as languages, shortcuts, and the address, model names and fetched model lists of AI services you configure.
- API keys, stored in the system Keychain and accessible only to Ferry and its system extensions.
- Translation history and favorites, including source/target text and thumbnails of pasted or shared images. You can turn history off, keep it forever, 90 days or 30 days, or clear it in Settings.
- Original audio from Conversation mode, kept only on your device, excluded from device backups, and deleted automatically after 7 days unless you favorite the conversation.
- Temporary recording frames and sound. When you record the screen on iPhone or iPad to make a long screenshot or a subtitle image, the key frames (and the sound, if you make subtitles from sound) are kept briefly on your device and deleted once Ferry has made the image; if you don't return to Ferry after recording, they are removed the next time you record. On Mac, frames are kept only in memory and recorded sound is deleted right after the lines are recognized.
- Speech cache. To save on your AI costs, audio that has been read aloud is cached on your device for a while and then deleted automatically.
We have no access to any of this data. Deleting the app deletes it from your device.
iCloud sync
When you use iCloud sync, your settings and translation history are stored in your own iCloud account and synced between the iPhone, iPad and Mac signed in to the same Apple Account. API keys and original audio from Conversation mode are never synced and stay on each device.
This data is held by Apple under its privacy policy, and we have no access to it. You can delete Ferry's data from iCloud storage management in system settings.
Permissions
- Camera (iPhone, iPad): for camera translation.
- Photos (Add Photos Only) (iPhone, iPad): only to save stitched long screenshots or bilingual subtitle images to your photo library. Selecting photos or videos uses the system picker, which requires no full library access.
- Screen Recording (iPhone, iPad): used only when you start a recording yourself to make a scrolling screenshot or a subtitle image, and paused while Ferry itself is on screen. The picture and sound are used only to stitch the image, recognize the lines and pick the matching frames, and are deleted once the image is made.
- Local Network (iPhone, iPad): only when you connect to a model service running on your local network, such as Ollama.
- Microphone and Speech Recognition: for voice input, audio transcription, Conversation mode, and when you choose to make a subtitle image from microphone sound. While the Conversation screen is open, the microphone stays on so the start of what you say isn't cut off, but only audio captured while you hold the button is recognized and saved. It turns off as soon as you leave the screen or switch apps. Where the device supports it, speech recognition runs on the device.
- Accessibility (Mac): only to read the text you select, for selection translate.
- Screen & System Audio Recording (Mac): only to capture the screen area you select (screenshots, screenshot translate, scrolling capture and Subtitle Stitch), and, when you start making subtitles from sound, to record the sound your Mac plays so the lines can be recognized.
Each permission is requested only when you first use the related feature, and you can turn it off at any time in system settings.
Exporting settings & general terms
You can export your settings to a file encrypted with a password you choose, to move them to another device. You keep the file; we never receive it.
Children
Ferry is not directed at any particular age group and does not collect personal data from anyone, including children.
Changes
If this policy changes, we will update this page and the effective date above. If we ever add a feature that collects data, we will tell you clearly before you use it.
Contact
Questions about this policy? Email service@huaguomt.com.