Privacy Policy

This policy applies to the hosted Jaime service at askjaime.xyz, the Jaime desktop app when it is signed in to a hosted workspace, and the private-beta waitlist. It is operated by Jaime (“we”). The free desktop edition running with your own provider keys and no account stores everything on your machine and sends nothing to us; only the provider you configured receives your prompts.

Things you give us

• Account: email address and password (hashed by our authentication provider), or the identity returned by Google or Microsoft when you sign in with them. A display name and avatar if you add one.
• Your content: conversations, projects, tasks, calendar events, documents and their versions, uploaded files and generated media, goals, habits, personal-life entries, notes, memories you confirm, agents and automations you create, and anything you paste or attach as context.
• Connected accounts: when you connect Google Calendar, Microsoft Outlook, Apple iCloud, or an MCP server, the OAuth tokens or app-specific passwords needed to sync, stored encrypted with a key that never leaves our servers. We sync the data you enabled (for example calendar events) and nothing else.
• Mailbox: if you connect an email account by IMAP, the password or app password you enter is stored encrypted and used only to read your inbox and send replies you approve. We keep a working copy of recent inbox messages (sender, subject, body text; no attachments) so they can be sorted and searched; the mailbox itself stays the system of record. Message text is sent to the chat model for triage and for drafts you request. Disconnecting deletes the credential and the copies.
• Billing: your plan, invoices, and payment status. Card numbers go to Stripe and never touch our servers; we store only Stripe’s identifiers.
• Waitlist: the email address you submit and the time you submitted it.

Things generated by using Jaime

• Usage records: a ledger of credits reserved, charged, and granted, with the model and modality of each generation, so we can show you where your credits went and correct mistakes.
• Receipts and audit trails: which model answered, which sources a research answer cited, what an agent did, and what you approved. These exist so you can review what the AI did on your behalf.
• Operational logs: request identifiers, error messages, and timing. We redact prompts and content from these; a log line tells us a request failed, not what you asked.
• Product analytics: page views with path, referrer, country, device class, and browser. No advertising identifiers, no cross-site tracking, no third-party ad networks.

When you connect Google Calendar, we read events and create or update events you request. When you connect YouTube, we read your channel’s videos and analytics and upload videos you review; uploads are private. Relevant imported history may be sent to the configured AI providers as bounded context for your requests. We do not use Google or YouTube account data to train generalized AI models.

Jaime’s use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy, including Limited Use requirements. You can disconnect in Integrations, or revoke access through Google account permissions. Disconnecting deletes our imported connector history without changing data in your Google or YouTube account. YouTube metadata is refreshed or removed within 30 days; authorization is checked daily, and confirmed revocation removes imported history.

Using the YouTube connection also involves the YouTube Terms of Service and Google Privacy Policy.

Jaime does not train models on your content. Each request is processed by the provider named on the Subprocessors page under that provider’s terms, which we link there. When you ask Jaime something, the prompt plus the context you attached — and only that — is sent to the model provider for that request. With screen consent, a question can include a screenshot. Relevant approved memory and connected-account history may also be included. The Context Inspector shows the exact context sent with each turn.

• Chat, image, video, audio and 3D generation go through our model gateway, PoYo, which routes to the model you chose or the one Jaime selected for you. The routing receipt names the model. Compatible chat and selected image/video requests may be sent to ToAPIs when the primary gateway is busy or unavailable.
• Search and document indexing send the title and plain text of documents you save, and the text of your search queries, to Jina to turn them into embeddings. Formatting, attachments, and version history are not sent. Deleting a document deletes its indexed chunks.
• Web research sends your query to a search provider and fetches public pages; short excerpts of those pages are passed to the model and stored with the citation.
• Meeting recordings you make or upload in the AI Note Taker go to the speech-to-text provider for transcription and to the chat model for notes. Recording requires you to confirm that participants were told and that recording is lawful where you are; that confirmation is stored with the meeting. Transcripts, notes and the audio are private to you. Nothing from a meeting becomes a task, event or memory until you create it from the review screen.
• Desktop dictation and selection actions capture only the field or selection you deliberately invoked Jaime on. Password fields, private windows, terminals, and apps you exclude are never captured.
• Agents act only within the permissions you granted and stop for approval before any side effect marked as requiring one. Every run is recorded in the trust ledger you can inspect.

The full list of providers, what each receives, and where it runs is on the Subprocessors & data flow page.

• Providing Jaime: storing, syncing, searching and generating the things you asked for.
• Keeping your account secure: authentication, rate limiting, abuse prevention.
• Billing, credit accounting, and answering support requests.
• Understanding aggregate usage so we can fix what is broken and decide what to build.
• Meeting legal obligations.

We do not sell your data. We do not share it with advertisers. We do not use your content to train models or to build profiles of you for anyone else.

Only the subprocessors listed on the Subprocessors page, each for the narrow purpose named there; people you explicitly share with (workspace members you invite, or anyone holding a share link you create); and authorities when the law requires it, in which case we will tell you unless we are prohibited from doing so.

Inside a shared workspace, members see what their role allows. Private projects and personal-life entries stay private to you even inside a shared workspace. When you remove someone from a workspace, their access ends immediately.

• Your content stays for as long as your account exists. Deleting an item removes it and everything derived from it (indexed chunks, embeddings) from the live database.
• Credit and billing ledgers are append-only while your account exists, because they are the record of what you were charged. They are deleted with the account, subject to the tax-record retention Stripe applies to invoices.
• Operational logs are kept for 30 days.
• Database backups retain deleted data for up to 30 days in daily snapshots. Encrypted archive copies used only for disaster recovery are kept for up to 12 months and are never restored selectively to recover one account’s data.
• Waitlist emails are deleted when you create an account or when you ask us to.

• Export: Settings › Export produces a complete download of every table you own rows in. It runs immediately, from the app, with no request to us.
• Delete: Settings › Account › Delete account removes your account, workspaces you solely own, your content, connected-account tokens, and stored files, and cancels any subscription. It runs immediately and cannot be undone. Shared workspaces you do not own keep the content you contributed to them, attributed to a removed member. Jaime and Cliq share one account, so deleting it in Jaime also deletes your Cliq account.
• Correct: everything you entered is editable in place.
• Disconnect: Settings › Integrations revokes any connected account and deletes its token.
• Ask us: for anything the app cannot do, or to exercise a right under GDPR, UK GDPR, CCPA/CPRA or similar law — access, portability, objection, restriction — email privacy@askjaime.xyz. We answer within 30 days. You may also complain to your local data-protection authority.

We process your account and content to perform our contract with you; security, abuse prevention and analytics under our legitimate interest in running a reliable service, which you may object to; billing records under legal obligation; and anything optional — connected accounts, desktop capture, cloud processing of a selection — with your consent, which you can withdraw by disconnecting or turning the feature off.

Data is encrypted in transit and at rest. Every table is protected by row-level security so one account cannot read another’s rows even if application code has a bug. Connected-account credentials are encrypted with a server-side key. Access to production is limited to the people who operate it and is logged. If we discover a breach affecting your data we will notify you without undue delay and, where the law requires, within 72 hours.

Our database and file storage run in the regions named on the Subprocessors page. Model providers may process a request in another region for the duration of that request; they do not retain it beyond what that page states. Where data leaves the EEA or UK we rely on the provider’s standard contractual clauses.

Jaime is not for anyone under 16. We do not knowingly collect data from children; if you believe we have, tell us and we will delete it.

When this policy changes in substance we update the date at the top and tell signed-in users in the app before the change takes effect. Continuing to use Jaime after that date means you accept the new version.