Privacy Policy
Effective 5 October 2026 · Last updated 5 October 2026
Summary
Researcher KMS is a field tool for documenting regional languages. It is used by language researchers, not by the general public. We therefore hold two kinds of data: account data for researchers, and language data from the speakers they interview.
- We never sell anyone's data and never use it for advertising.
- Speaker recordings live on private storage, never in an open public space.
- Speaker permission is recorded — verbal or written, together with the scope of use.
- You can ask us to delete your data; see “Your rights”.
1. Who runs this service
The data controller is Peter Kambey, a researcher and the owner of this language documentation project (an individual project, not a registered company). The service runs at kms.idlce.ai.id. Contact details are in section 10.
2. What we collect
| Type | Contents |
|---|---|
| Researcher account | Name, email address, optional messaging number, and device push token. |
| Speaker records | Name, age, gender, village, sub-ethnic group, occupation, field notes, and the consent record (status, form, date, scope, notes). |
| Recordings and transcripts | Interview audio files, interview date and location, and the text segments produced from those recordings. |
| Questionnaire answers | Word-list (Swadesh) answers and the researcher's annotations. |
| System records | Change history (who edited or removed which entry) and technical error logs. |
The app does not request GPS location, contacts, calendar, SMS, or any file you did not pick yourself. The only device permissions we use are INTERNET and POST_NOTIFICATIONS.
3. Why we collect it
For research and preservation of the language: building a dictionary, a grammar, and a text corpus. The data serves research and a free public service — not advertising, not resale, and not profiling.
4. Speaker consent
Before recording or collecting anyone's language data, the researcher asks the speaker for permission and records it in the app: whether permission was given or refused, in what form (verbal or written), and the scope of use — whether the recording may be kept, whether the speaker's name may be shown, and whether the material may be published.
Where a speaker has refused, we do not serve their recording and we exclude their name from anything we share. Older speaker records whose permission was never captured are left untouched and are not published until we have asked again.
5. Who we share it with
We do not sell or rent data. A few third parties process data on our behalf, only as far as the service requires:
- Language-model provider (ai.paas.id) — receives the text you submit for translation or analysis. Speaker names are not part of that text.
- Google (Firebase Cloud Messaging) — delivers notifications to a researcher's signed-in device.
- Backup storage (Cloudflare R2) — encrypted copies kept for disaster recovery, so the language record is not lost.
- Messaging platforms (Telegram, Facebook, WhatsApp) — only if you choose to use the public service through those platforms.
Speaker recordings are never sent to the language-model provider.
6. How long we keep it
- Speaker records and recordings — kept for the duration of the research and not deleted automatically. We deliberately did not enable automatic deletion of interview recordings, because a recording cannot be recreated.
- Unreferenced files (pointed to by no record) — moved to quarantine and removed after 30 days.
- Daily backup copies — kept for 3 days, then replaced by a newer copy.
- Audit history — kept without a time limit, because a deleted audit trail can no longer tell what happened.
7. Your rights
- Researchers (account holders): you can delete your own account inside the app (Account → Delete Account). That permanently removes your account and the recordings, transcripts, sessions, speaker records, and answers you created. The system audit trail is retained.
- Speakers (interviewees): you may ask for a copy of your data, ask us to correct it, or ask us to delete it. Write to the contact in section 10; we will act on it and tell you what was removed.
8. Security
All traffic uses HTTPS. Interview recordings are stored on private storage reachable only by authorised accounts, and every link to a recording requires authentication. Accounts are created by an administrator only — there is no self-service sign-up. If data is ever exposed, we commit to notifying the people affected rather than keeping the incident to ourselves.
9. Children
This app is intended for adult language researchers. It is not directed at children and not intended for anyone under 13.
10. Contact
Questions, requests for a copy, or deletion requests: [email protected]. We aim to respond within 7 days.
11. Changes to this policy
If this policy changes — for example once a retention period for recordings is decided — we update the date above and publish the current version on this page.