Who we are
Kodros publishes Homelab Dictate. Tanner Osterkamp runs Kodros. In this policy, "Kodros", "we" and "us" mean Kodros. Questions go to tanner@kodros.ai.
What we collect
Nothing. The app sends nothing to Kodros. It contains no analytics, advertising or crash reporting code.
Your voice
- When it listens. The microphone opens when you start a session. iOS shows its microphone indicator while it is open. Nothing is recorded until you tap the mic button or the dictate key. The session ends after the time you choose, from 1 to 60 minutes, or when you tap End session.
- What is recorded. Each dictation makes a short recording in the app's temporary folder. The app deletes it when the transcription finishes or fails.
- Without your own server. The app uses Apple speech recognition. When your iPhone supports your language on the device, recognition runs on the iPhone. Otherwise the recording goes to Apple's speech recognition service, and the main screen says so before you record. Apple handles it under its own privacy policy.
- With your own server. The recording goes in one request to the address you entered, with your access token. It goes nowhere else. The request uses HTTPS, or plain HTTP only to an address on your local network. Whoever runs that server decides what it keeps.
What stays on your iPhone
- Server address. Kept in the app's settings.
- Access token. Kept in the iOS keychain on this iPhone, in a group only the app can read. The keyboard never receives it.
- Latest transcript and status. The app passes the newest transcript, any error and a small status record to the keyboard through a keychain group shared by the app and keyboard. The next transcript replaces it. Ending the session clears it.
The keyboard and Full Access
The keyboard asks for Allow Full Access because iOS requires it before a keyboard can read the shared keychain group. The keyboard has no network code and does not record what you type. It reads the few characters before the cursor only to decide whether to add a space.
The demo
Try the demo types sample sentences. It does not use the microphone, contact a server or save an address.
Permissions
- Microphone. To record while you dictate.
- Speech recognition. To use Apple speech recognition when you have no server set up.
- Local network. iOS may ask the first time the app reaches a server on your home network.
- Allow Full Access. For the keyboard, as described above.
Retention
We hold no data from Homelab Dictate, so we keep nothing. Recordings are deleted after each transcription. Your server and Apple keep what their own policies say.
Your choices and deletion
Disconnect removes the saved address and token. Deleting the app removes its settings. iOS may keep keychain items after deletion.
To ask what we hold about you, or to ask us to correct or delete it, email tanner@kodros.ai. We answer by email.
Children
Homelab Dictate is not directed to children. We do not knowingly collect personal information from children under 13. If you think a child sent us information, email us and we will delete it.
Changes
If Homelab Dictate changes how it handles data, we will update this page and the effective date at the top. If a change is significant, we will say so at the top of the page.
Contact
Email tanner@kodros.ai with any question about this policy.