Privacy Policy
Effective October 28, 2026. Ember is a journaling app, so almost everything in it is something you wrote. This page explains where that writing goes, who can reach it, and how to take it back. It describes what the app actually does, not what we wish it did.
Who is responsible
Ember is operated by Avraham Meir Shore, a sole proprietor based in Toronto, Ontario, Canada. For any question about your data, including requests to see, correct, or delete it, write to [email protected].
The short version
- Ember collects your email, what you write, and what the app generates from what you write.
- There is no advertising, no tracking, no analytics product, and nothing is sold or shared for marketing.
- Your entries are encrypted at rest. This is server-side encryption, not end-to-end encryption.
- AI features send the relevant parts of your writing to an AI provider to produce the result.
- You can export everything or delete your account from inside the app, at any time.
What Ember collects
Your account. Your email address, and either a password or a Google sign-in, handled by our authentication provider. Passwords are stored hashed, never in plain text, and we never see them. Optionally a display name, an avatar, and your settings. We use your email address to sign you in, reset your password, send security or policy notices, and reply to messages you send us — never for marketing.
What you write. Journal entries and drafts, imported entries awaiting review, highlighted passages, habits and their daily logs, mood and energy check-ins, saved insights and notes, quotes you keep, sticky notes, and any feedback you send us.
What you upload. Photos of handwritten pages and documents you import, which are sent to an AI provider to be transcribed.
What Ember generates about you. Entry summaries, detected themes and emotions, weekly and monthly summaries, longer-term pattern analysis, habit correlations, and suggested prompts. These are derived from your writing and are as personal as the writing itself.
Operational records. For each AI request: which feature ran, which model, how many tokens, and how long it took. These records contain no text from your entries. Also which dashboard quotes were shown to you and when. Your IP address is used briefly to rate-limit abusive traffic, and appears in our hosting providers' server logs.
Camera and photo access
Ember's handwriting scan opens your device's camera or photo library through the standard system picker, so you can photograph a written page or choose an existing image. Ember reaches for that access only when you start a scan, and you can withdraw it at any time in your device settings. The image is sent to an AI provider to be transcribed; we keep the resulting text, not the photo.
What Ember does not collect
No location data, no contacts, no browsing history outside the app, no payment information, no advertising identifiers. Ember contains no third-party analytics, attribution, or crash-reporting SDKs, and does not track you across other apps or websites. We do not sell your data, and we do not share it with anyone for their own purposes.
How your writing is stored
The title and body of your journal entries, drafts, and pending imports are encrypted in our database, and access is restricted to your account. This is server-side encryption: it is not end-to-end encryption. Ember holds the key and decrypts your text whenever a feature needs it, which means the operator has the technical ability to read your entries.
The material Ember derives from your entries is stored unencrypted. That includes entry summaries, weekly and monthly reflections, themes and emotions, longer-term pattern analysis, the history of every AI output, passages you highlight, quotes and focus cards you keep, generated prompts, notes on your habits, questions you ask about your own journal, and mood and energy numbers. Some of it is your writing word for word — a highlight or a saved quote stores the passage you selected exactly as you wrote it.
It is worth being precise about what the encryption above does and does not buy you, because the difference is smaller than the word usually suggests. Both your encrypted entries and the unencrypted material derived from them are protected by the same thing: access rules enforced inside the database, per account, so that no other user can reach your rows. Encryption adds one layer on top of that. If someone obtained a copy of the database itself — a stolen backup, a leaked dump — your entry text would be unreadable to them without the separate key, while the derived material would be readable. Against us, or against a compromise of the servers that run Ember, encryption protects neither, because the key and the data live in the same system.
So the honest statement is this: Ember is built so that no other user can reach your journal, and we do not read your entries — not to improve the product, not out of curiosity. We read your content only if you ask us to for support, or if we have to in order to fix a fault or meet a legal obligation. That is a commitment about our conduct, not a guarantee enforced by cryptography. If you need a journal that is mathematically unreadable to its operator, Ember is not that product today.
Sharing an entry
Ember lets you share a single entry by turning on Spark for it. Doing that creates a secret link, and anyone who has that link can read the full entry without signing in or holding an Ember account. The link is the only thing protecting it: there is no password, and we do not track who opens it. Nothing is shared unless you switch Spark on for that specific entry.
Turning Spark off again destroys the link immediately and the entry stops being reachable; a link you shared previously will no longer work. Links do not expire on their own, so an entry you sparked stays readable to anyone holding the link until you turn it off. If you have shared something you would rather not have, turning Spark off is the way to take it back.
AI features and the providers behind them
When you ask Ember to reflect on an entry, find patterns, transcribe handwriting, or generate a prompt, the relevant content is sent to an AI provider, which processes it and returns a result. Depending on the feature and the model selected in Settings, that provider is Google (Gemini, the default) or OpenAI.
None of this happens until you agree to it. The app asks for that permission on a consent screen before anything is sent, and the choice stays yours: Settings → Ember → AI features turns AI off or back on at any time. With it off, our servers refuse every AI request for your account — including ones you trigger yourself — and Ember works as a plain journal and habit tracker.
Your content is sent for the sole purpose of producing the output you asked for. We use these providers through their paid APIs, whose terms state that content submitted this way is not used to train their models. Providers may retain requests for a limited period for abuse and safety monitoring, under their own policies:
Ember never sends your content to a provider for any purpose other than the feature that asked for it. You should know that not every one of those features waits for you to press something. Saving an entry sends it to be analysed so Ember can show you what it noticed. Opening the dashboard or the insights pages can generate a reflection, a prompt, a weekly summary, or a set of habit observations if there is not already a recent one stored. Searching your journal in more than a few words sends the query to be interpreted. These run on your behalf, to fill the parts of the app you are looking at, and they are the reason Ember feels ready when you arrive.
You can switch all of it off. Settings → Ember → Generate automatically stops every one of those unprompted calls. With it off, saving an entry no longer sends it anywhere, the dashboard and insights pages stop generating on their own, and searching no longer sends your query to be interpreted. Nothing reaches an AI provider until you ask for it by name, and every feature still works when you do — the insight on an entry, the reflection, the weekly summary, each one a button away. Anything generated before you switched it off stays where it is.
That switch is enforced on our servers, not merely respected by the app on your device, so turning it off is a decision the service honours rather than a preference the client is trusted with. It covers everything unprompted, including the optional typo cleanup and title suggestions that run on save — if you had those turned on, switching this off stops them too. You also keep the finer controls: the refresh settings slow individual features to manual, and cleanup and titling stay off unless you turn them on.
If you connect Ember to Claude using the integration in Settings, Claude can read the journal data you authorize it to read, and that data travels to Anthropic as part of your Claude conversation. You grant that access, and you can revoke it at any time in Settings.
Who else handles your data
- Supabase — database, authentication, file storage, and the server functions behind every feature. Your data is stored in the United States (Oregon).
- Vercel — hosts and serves the app itself.
- Google, Anthropic, OpenAI — AI processing, as described above.
These providers act on our behalf, under their own security and privacy commitments. We add no one else without updating this page.
Where your data lives
Ember is operated from Canada, and your data is stored and processed in the United States. AI providers may process your content in other countries. By using Ember you accept that your information crosses borders this way.
How long we keep it
Your content stays until you remove it. An entry you move to Trash stops feeding reflections and analytics straight away, stays recoverable in Settings for 30 days, and is then deleted automatically along with the analysis Ember generated for it. You can restore it or delete it yourself at any point in those 30 days.
One part of that deserves to be said plainly, because it can surprise people. Anything you deliberately kept out of an entry survives the entry itself. A passage you highlighted, a quote you saved to your library, a focus card you pinned, and the stored history of AI output all remain after the entry they came from is gone, and the excerpts among them are your words verbatim. That is deliberate — you kept them on purpose, and we do not think deleting an entry should silently take your saved quotes with it — but it does mean deleting an entry is not by itself enough to erase every copy of what it said. Highlights, saved quotes, and focus cards can each be deleted directly where you keep them. Stored AI output history currently cannot be cleared on its own — today it goes only when you delete your account, and we are adding a way to clear it from Settings. Deleting your account removes all of it.
Pending imports stay until you import them, discard them, or delete your account. When you delete your account, your rows are erased from our live database immediately and together, including AI output history, saved insights, integration tokens, and sessions. Encrypted database backups may retain copies for up to 30 days before they age out. Per-request operational records — which feature ran, which model, token counts and timing, never any journal text — may be retained after account deletion for service accounting.
Your controls
- Export. Settings → Journal storage → Export downloads a ZIP of your entries, habits, check-ins, insights, quotes and focus cards, in both JSON and CSV.
- Delete. Settings → Journal storage → Delete Account removes your account and your content.
- Correct. Edit or trash any entry, habit, or insight directly in the app.
- Revoke. Disconnect the Claude integration, or remove your own AI key, in Settings.
Depending on where you live, you may also have the right to access, correct, delete, or port your data, to withdraw consent, to object to processing, and to complain to a privacy regulator. Under Canadian privacy law, including PIPEDA and Quebec's Law 25, you can ask us what personal information we hold about you, how it is used, and who it has been shared with.
If you are in the European Economic Area or the United Kingdom, our legal bases for processing are: performing our contract with you (providing the journal you signed up for), our legitimate interests (keeping the service secure and preventing abuse), and your consent where required. You can withdraw consent at any time, and you can lodge a complaint with your local data protection authority.
The person accountable for privacy at Ember is Avraham Meir Shore. Write to [email protected] and we will respond within 30 days. If our answer does not satisfy you, you can contact the Office of the Privacy Commissioner of Canada.
Sessions and local storage
Ember stores a sign-in token and your interface preferences in your browser or device storage, so you stay signed in and the app looks the way you left it. These are necessary for the app to work. Ember sets no advertising or cross-site tracking cookies.
Children
Ember is not intended for anyone under 16, and we do not knowingly collect information from children under 16. If you believe a child has created an account, write to us and we will delete it.
Security
Access to your rows is enforced at the database level, per account. Entries are encrypted at rest, traffic is encrypted in transit, and integration tokens are single-use and rotate. No system is perfectly secure; if a breach ever affects your data, we will tell you and the appropriate regulator without undue delay.
Changes to this policy
If we change how Ember handles your data, we will update this page and move the effective date above. For changes that materially affect you, we will also tell you in the app or by email.
Contact
Questions, requests, or concerns: [email protected]. For help using the app, see Support.