Privacy policy
Last updated 10 November 2026. This Diario is operated by the operator of this Diario deployment.
Diario stores what it needs to run your Spanish practice and nothing else. It does not sell data, show ads, or use your writing to train models.
What Diario collects
- Account. Your email address, and your name if you sign in with Google.
- Practice. The texts you write or say, their corrections and scores, the errors your tutor tags, dictation attempts, drill answers, your review deck, session summaries, streak and XP.
- Settings. Time zone, session length, and your choices about register, region and explicit content.
- Connections. Which apps you connected and when. Access tokens are stored only as hashes.
- Sign-in protection. Your IP address is used briefly to limit repeated failed sign-ins. The hosting provider also keeps standard request logs.
- In your browser. A cookie keeps you signed in. On the practice page, another remembers which source you asked to see, and the browser's own storage holds your unsent draft and your choice of explanation language. None of these is used to track you.
What Diario does not collect
- Your conversations with your assistant. Diario only sees what the assistant sends to a Diario tool: for example, the text you wrote and its correction.
- Your assistant's memory, chat history or files.
- Recordings of your voice. Diario never receives or stores audio from you.
How it is used
Only to provide the practice: choosing what to review, scoring dictations, tracking progress and showing it back to you. It is not used for advertising, profiling or model training, and it is not sold or shared for anyone else's purposes.
Who else handles data
- Hosting and database. Vercel hosts the service and clip audio; a managed Postgres provider stores the database.
- Google, only if you choose Google sign-in, to confirm your email address.
- Deepgram transcribes public podcast audio to make the listening clips. It receives no learner data.
- OpenAI, only if you send a text for correction or ask for a passage to be explained on the practice page. For a correction Diario sends OpenAI the text you wrote, the task, and the opening of the article you wrote about, and receives the correction. For an explanation it sends the words you selected and the paragraph around them. Your name and email are not sent. Diario asks OpenAI not to keep the exchange as a stored response; OpenAI's API terms govern anything else it retains.
- OpenAI, only if you use live speaking practice. Your microphone audio goes from your browser directly to OpenAI, which returns the spoken replies and a transcript. Diario stores the transcript text.
- Your assistant's provider (for example Anthropic or Meta) receives the results of Diario's tools as part of your conversation, under its own terms and privacy policy.
How long it is kept
- Practice data and settings: until you delete them or your account.
- Listening clip audio: about 45 days, then deleted.
- Access tokens: one hour. Refresh tokens: thirty days from last use.
- Sign-in throttling records: fifteen minutes.
Your controls
- Export everything stored about you from your dashboard.
- Delete your learning data, or your whole account, from the dashboard or by asking your tutor to delete your data.
- Disconnect any app, or revoke any personal token, from the dashboard.
Security
Connections use HTTPS. Sign-in to the connector uses OAuth 2.1 with PKCE. Tokens are random, stored only as hashes, and can be revoked at any time. If a security incident affects your data, the operator will tell you without undue delay.
Age
Diario is intended for adults. It is not directed at anyone under 18.
Changes and contact
If this policy changes in a way that matters, the date above changes and signed-in users are told on the dashboard. For questions or requests, contact the operator of this Diario.