ANIS Privacy Policy

Effective date: 24 October 2026
Policy version: 2026-10
Operator: Mohamed Maslooh (محمد مصلوح), the developer and publisher of ANIS (“we”, “us”)
Contact: contact@maslooh.com

This policy explains what data ANIS processes, why, and the choices and rights you have. It applies to the ANIS desktop application (Windows), the ANIS Android application, and the ANIS website at anis.maslooh.com (including its browser trial, served from anis.adamax.online/try). The Arabic version of this policy is published at https://anis.maslooh.com/privacy; in case of doubt about interpretation, contact us at the address above.

1. What ANIS does

ANIS is a voice transcription product. You record audio (or upload an audio file); ANIS converts it to text using an AI model and delivers the text where you are working. ANIS is designed to be local-first: your transcript history and recordings live on your device, not on our servers.

2. Data we process on our servers (signed-in use)

Account data: your email address and Google account identifier (sub claim), obtained when you sign in with Google; your subscription plan; account creation date; and your model-improvement consent choice (the choice itself, when it was made, and which policy version you saw).
Usage metadata per transcription request: recording duration, word count, model token counts, provider cost, processing latency, the Command used, whether the request was dictation or a file upload (they are metered separately, see the Terms section 3), and a timestamp. This is used for quota metering, fair-use monitoring, billing, abuse prevention, and service monitoring.
Audio recordings: sent to our server to be transcribed. Audio is forwarded to our AI provider (section 5) to produce the transcript and is not kept by ANIS after the request completes, except (a) as described under “Model-improvement data” (section 3) when you have agreed to it, and (b) transient operational logs at our infrastructure and AI-gateway layer, which may hold request payloads for a limited period (no longer than 30 days) solely for troubleshooting, abuse prevention, and service monitoring. If you enable the optional system-audio capture setting on Windows (off by default), your recording also includes audio playing through your device (for example the other side of a call), and that audio is processed exactly like microphone audio. You are responsible for complying with the recording-consent laws that apply to you.
Transcript text: returned to your device. ANIS servers do not keep your transcript text after the request completes, except as described in section 3 (model-improvement data, with your consent), section 4 (personalization requests and reports/feedback you send us), and the transient operational logs described above.

If you have agreed on the in-app consent screen, we keep a copy of each transcription request to improve ANIS’s transcription quality, including training and evaluating the AI models that power ANIS:

the audio recording,
the resulting transcript text,
the instructions the AI model was given for that request,
and request metadata: the Command used, the name and category of the application you were dictating into, whether smart formatting adapted the wording for that application, the language you translated into, recording duration, the model that served the request, whether the request was a retry, your app version, platform, and interface language, a timestamp, your account identifier, and the consent version you accepted.

We keep the application name and the smart-formatting state because with smart formatting on, ANIS does not transcribe you word for word: it adapts the wording to the program you are dictating into. Without knowing which program that was and that the adaptation happened, the saved transcript cannot be interpreted correctly and is useless for improving the models. Only the application’s name and category are recorded, never its contents, window titles, or anything you have open in it.

Recordings transcribed on your phone. When the Android app transcribes a dictation on the phone itself (Settings → Local model), the audio is not sent to us to be transcribed. If you have agreed (consent version 2026-10 or later), the app keeps a copy of each such recording and uploads it later in the background, with: the text the on-device model wrote, the text you kept after any edits or applied suggestions, the text of each part of the recording with its start and end time, the version of the on-device model, the recording duration, your app version, platform and interface language, your account identifier, a timestamp, and the consent version you accepted. Files you import into the app (lectures, voice notes, videos) are never uploaded under this program, because they can contain other people’s voices.

This data is stored securely on our infrastructure (Cloudflare R2, see section 5). It is never collected from users who have not agreed, and never sold or shared with advertisers. Your consent is recorded with a timestamp and policy version.

You can withdraw at any time in the app (Settings → Data sharing). Withdrawing stops collection immediately and deletes everything previously collected from you (completed within 30 days at the latest). Deleting your account has the same effect.

4. Other data you may send us

Personalization (if enabled): to learn your writing style and custom vocabulary, the app periodically sends a set of your recent transcripts, together with your current personalization rules, to our server, which passes them to an AI model and returns updated rules or a style profile to your device. These requests are processed transiently: the server stores neither the transcripts nor the result; the result is saved only on your device.
Transcription reports: if you report a transcription from inside the app, the reported transcript text, the report category, and any note you enter are sent to us for quality review. The audio is not attached to a report.
Feedback: if you use “Send feedback” in the app, we receive your message, the feedback category, and basic context (app version and platform) so we can act on it.

5. Third-party processors

We use a small number of service providers to run ANIS. They process data on our behalf and under their own contractual and legal obligations; we do not allow them to use your data for their own advertising.

Processor	Purpose	Data shared
Google (OAuth)	Sign-in	OAuth identity (email, subject id)
Cloudflare	Hosting (Workers, D1), storage of model-improvement data (R2), AI Gateway	All server-side data above
OpenRouter and its underlying AI-model providers (currently Google Gemini for transcription; additional text-model providers for auxiliary features such as personalization)	AI transcription and text processing	Audio and/or text + the processing prompt
Lemon Squeezy	Payments and subscription management (merchant of record)	Email, plan, payment data (ANIS never sees card numbers)
GitHub	Distribution of app downloads and updates	Standard web-request data (e.g. IP address) when the app checks for or downloads an update

Our AI providers process your audio and text only to return the requested output. Under the terms governing our paid API usage, they do not use your content to train their models.

6. What we do NOT do

We do not sell your data.
We do not use your audio or transcripts for advertising, and we do not share them with advertisers or data brokers.
We never use any user’s audio or transcripts to train AI models or for model improvement without the consent described in section 3.
We do not show ads.
We do not read the contents, window titles, or screen of the applications you dictate into; only the active application’s name and category are used, to adapt formatting.

7. Legal basis

Where data-protection law applies (including the EU/UK GDPR, the Saudi Personal Data Protection Law (PDPL), and Egypt’s Personal Data Protection Law No. 151/2020), we process:

account data, audio submitted for transcription, and usage metadata as necessary to provide the contracted service (transcribing your audio, metering quota, billing);
security logging, abuse prevention, and service monitoring on the basis of our legitimate interest in keeping the service safe and working;
model-improvement data (section 3) on the basis of your consent, given by an affirmative action on the in-app consent screen. Consent is never assumed, is recorded with a timestamp and policy version, and can be withdrawn at any time with the effects described in section 3 (collection stops and previously collected data is deleted).

8. Retention

Account and usage metadata: kept while your account exists.
Audio and transcripts outside the model-improvement program: not retained after transcription completes, other than transient operational logs (≤ 30 days, section 2).
Model-improvement data: kept while your consent stands; deleted when you withdraw consent or delete your account, completed within 30 days at the latest.
Reports and feedback: kept as long as needed for quality review and product improvement.
You can delete your account at any time from inside the app (Settings → Account → Delete account). Deletion is immediate and irreversible: your account record and all server-side data keyed to it (usage metadata, quota state, sign-in sessions, and any model-improvement data) are permanently deleted when you confirm. Data stored on your device is not affected.

9. Data stored only on your device

Transcript history (local SQLite database).
Audio archives of your recordings (in a folder you can choose).
Personalization data: your custom vocabulary, user rules, style profile, and custom Commands.
App settings, including the name/category of the application you were dictating into (used to adapt output formatting). This leaves your device only as part of the transcription request itself, and is stored on our servers only under the model-improvement program (section 3).
On Android, recordings transcribed on the phone by the Local model and their text. They leave the phone only if you agreed to share them (section 3).
Your session token, stored in the operating system’s secure keyring. In the browser trial on our website there is no such store, so the session token is held in your browser tab’s own storage and is discarded when you close the tab.
In the browser trial, the texts it produces for you, kept in your browser’s local storage so the trial can show you what you have already tried. They are never sent anywhere and we never hold a copy: clearing the list on the page, or clearing this site’s data in your browser, removes them.

You can delete on-device data at any time by clearing history in the app or uninstalling it.

10. Security

Data in transit is protected with TLS. Server-side data is stored with providers (Cloudflare) that encrypt data at rest. Your session token is kept in the operating system’s secure credential store, and access to server-side data is restricted to what operating the service requires. No system is perfectly secure; if we become aware of a breach affecting your personal data, we will notify you and the competent authorities as the law requires.

11. International transfers

Our infrastructure runs on a global cloud network, so your data may be processed outside your country of residence. Where transfer rules apply (e.g. GDPR, Saudi PDPL), we rely on our processors’ contractual safeguards and applicable transfer mechanisms.

12. Google user data

ANIS’s use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements. We use Google sign-in data (your email address and account identifier) only to create and operate your ANIS account.

13. Your rights

You can exercise the most important rights directly in the app:

Delete everything: Settings → Account → Delete account (section 8).
Withdraw model-improvement consent: Settings → Data sharing (section 3).

In addition, subject to applicable law (GDPR, Saudi PDPL, Egypt PDPL), you have the right to request access to, a copy of, correction of, deletion of, or restriction of the processing of your personal data, to object to processing based on legitimate interests, and to data portability. To exercise any of these, contact contact@maslooh.com; we respond within 30 days. If you believe we have not handled your data lawfully, you may lodge a complaint with the data-protection authority competent for you.

14. Children

ANIS is not directed at children under 13 (or the higher minimum age that applies in your country), and we do not knowingly collect their data. If you believe a child has provided us personal data, contact us and we will delete it.

15. Changes to this policy

We will post any changes to this policy at the URL above and update the effective date and policy version. Material changes will be announced in the app; a material change to the model-improvement program (section 3) re-triggers the consent screen before any further collection under the new terms.

16. Contact

Mohamed Maslooh (محمد مصلوح), developer and publisher of ANIS
Email: contact@maslooh.com

Privacy Policy