1. Scope and our privacy approach
FROMYOU, LLC, a Delaware limited liability company ("FROMYOU," "we," "us," or "our"), provides Stella's hosted services and official desktop, browser, and mobile applications. This Privacy Policy explains the personal information we process through those services, including agents, Drive, sharing, hosted apps and previews, and Stella Crew for Slack.
Stella is privacy-first, but it is not a local-only or entirely offline assistant. Our hosted service requires an account. It stores authoritative conversation and task records, including when work runs on your computer or uses your own model account or API key. Content relevant to a task can be sent to AI providers, connected services, or recipients you choose. We do not sell your personal information, use your conversations to serve advertising, or use your conversation content to train AI models. Our Google API commitments are stated in section 6.
Self-hosting
When you run Stella on your own server, you operate it and control its data; FROMYOU does not receive or process that server's data. These Terms and this policy cover only our hosted service and official apps, not your server. Official apps may still contact app stores or Expo for downloads and OTA updates, and, if you enable it, our push relay for notification delivery. Those services process the update or notification information needed to deliver them.
2. Information we collect and where it is stored
Account and sign-in information
For hosted accounts, we process your email address, name and profile information if supplied by you or an identity provider, account identifier, creation time, authentication records, and preferences. Available sign-in methods include email sign-in and supported identity providers such as Google and Apple; the web experience also supports passkeys. We use the Better Auth software for authentication. Stella's email/social sign-in does not require you to give us your Google or Apple password. Passkey authentication uses a public-key credential; your passkey's private key is managed by your authenticator, not sent to us.
Authentication and security records can include session identifiers, provider account identifiers, sign-in time, IP address, user-agent information, and device information. Signing in to Stella with Google is separate from granting access to Gmail, Drive, or other Google services.
Conversations, memory, and agent work
We store messages, replies, conversation and event history, saved memory, attachments and references, tool calls and results, agent status and activity, and settings needed to synchronize clients and continue work. Relevant content may include file excerpts, screenshots, web pages, email or calendar content, shell commands and output, generated code or media, voice transcripts, and other information you supply or direct agents to use. Engine session and continuation records, including Claude Code session transcripts and tool outputs, can also sync to Stella's hosted services so work can resume across execution locations.
Agents can read and change files, use applications and your signed-in browser, and act on connected accounts. We do not upload every file or browsing record simply because it exists on your computer. However, content an agent reads for a task may enter its model context, tool results, conversation, memory, or hosted workspace. Local execution is not a promise that the resulting information remains local. Public web requests also disclose information to the sites contacted.
The hosted service maintains authoritative records using Cloudflare infrastructure, including database and object storage. Supported clients keep local caches and execution state. Device-local information can include preferences, working files, screenshots, local memories, installed skills, credentials, browser state, and modification history. Deleting a cache or uninstalling an application does not delete the hosted record. Likewise, deleting hosted records does not automatically erase files or credentials remaining on your computer.
Hosted agents, Drive, and apps
Cloud agents run in hosted sandboxes. Files, dependencies, credentials needed for execution, browser state, working directories, and continuation/checkpoint data can be stored or restored between runs. A sandbox's lifecycle is not the retention period of your files or task history. Drive stores uploaded and generated files and metadata for cross-device use. Hosted apps, sites, and previews store the code, assets, deployment and access settings, and, where applicable, application data needed to serve them.
Device and routing information
We process device identifiers, device names and platform, public identity keys, connectivity and last-seen information, execution location and conversation mappings, permissions for cross-device work, and delivery or cancellation records. This lets us authenticate devices, route tasks, synchronize clients, and recover interrupted work. Mobile notifications use a push token and notification-related identifiers and metadata.
Connected accounts, schedules, and collaboration
We process authorization and connection records and the inputs, results, and selected content needed to use a service you connect, including external MCP servers. Ongoing features such as schedules, watches, and email summaries can access the connected service under your continuing instruction. Revoking a connection stops future authorized access but does not undo completed actions or remove all historical copies.
Stella Crew processes Slack installation, workspace, user, channel and thread identifiers, relevant messages, repository configuration and content, access and environment settings, agent instructions, results, and preview information. Slack messages, repository changes, and shared outputs may also remain with Slack, GitHub, the organization, or other participants. Other collaboration features, where used, can process profiles, memberships, messages, and shared-session activity.
Billing, usage, and support
Stripe processes hosted subscription payments. We process customer and subscription identifiers, plan and payment status, billing periods, invoices or billing references, and payment-method display information such as brand and last four digits where available. We do not store full payment-card numbers or CVVs. If store billing is offered, Apple or Google processes the transaction under its own policies.
We record usage and operational metadata such as account or owner identifier, model and engine, agent type, tokens, time, estimated or charged cost, plan, limits, success or failure, and request identifiers to operate the Service, enforce limits, reconcile billing, and prevent abuse. Support communications include the information you choose to send us. Technical and security logs may contain identifiers, request and error information needed to investigate incidents. These records are distinct from optional telemetry.
Optional discovery and information about other people
If you enable onboarding discovery or personalization, Stella can use selected information such as bookmarks, applications, and development-environment details to help personalize your experience. Discovery can be skipped. Sensitive discovery categories such as Messages & Notes are opt-in and disabled by default. Resulting context or memory may be processed by a model and synchronized to the hosted service; discovery is not a guarantee of device-only processing.
Files, messages, and connected accounts may contain other people's personal information. Use only information you are authorized to provide or process and follow any notice, consent, or organizational requirements that apply.
3. AI data sharing: who receives what
Using an AI feature sends the content needed for that feature to the selected AI service. Depending on the route and feature, recipients can include Anthropic, OpenAI, Google, xAI, DeepSeek, and Meta, and the AI gateways OpenRouter and Fireworks. User-configured routes can also include GitHub Copilot, Moonshot AI/Kimi, and Z.AI, when selected. A gateway may forward the request to its selected upstream model provider. Not every recipient receives every request.
The content can include your prompt, attachments and images, relevant conversation and memory, instructions, tool descriptions, file or screen content, tool results, audio, and model output. Stella's infrastructure can process these inputs and outputs to route work, synchronize clients, record history, and deliver the feature.
• Stella-managed models: requests go through our hosted infrastructure and the gateway or model provider used for that request. • Your API keys: supported direct-device routes send requests to the chosen provider without our managed inference relay. Account-configured keys and cloud routes can instead be stored in Stella's credential infrastructure and used through our backend or model gateway. BYOK does not always mean local storage or a direct-device request. • Claude and ChatGPT logins: the appropriate engine provider processes the session context and results. Credential availability depends on execution location. Conversation and engine-session records can still sync to Stella's hosted service. • Voice and dictation: audio and related text can be sent to Meta's transcription service, OpenAI's realtime voice service, or another provider you configure for the supported route. • Media generation and transcription: prompts, source files, audio, generated media, and job information can be sent to fal.ai and the selected media or transcription provider.
Our model relay may temporarily cache assembled response content, including text, reasoning, and tool arguments, for request recovery or reuse. These operational buffers are separate from encrypted conversation history and need not use the same application-level encryption. Retention follows the request and cleanup lifecycle rather than a universal ten-minute limit.
AI providers have their own retention, safety, and account policies. Some API routes use provider-side response storage or continuation features. Stella does not guarantee zero retention for every provider or route, nor does using your own key guarantee it. Check the selected provider's terms and available account controls. Our commitment not to train models on your conversation content does not substitute for a third party's separate policy; Google API data is additionally protected by section 6.
The mobile app presents an AI data-sharing disclosure and consent control when setting up AI use. You can decline and avoid those AI features. To stop sharing through a particular route, stop using it, change your engine, or disconnect it where supported. Changes do not recall data already sent. Consent to AI use is not consent to use Google API data for advertising or generalized model training.
4. Credentials and secure-input features
Stella stores credentials needed for the route you configure, which can include provider API keys, connected-service OAuth tokens, engine credentials, and MCP authorization. Some remain on the selected computer; others are held by hosted credential services for account-wide or cloud execution. Local credential protection uses the operating system's credential storage or other protected storage appropriate to the platform. Server-side credential vaults use encryption. A locally connected engine account is not necessarily available on every other device.
Secure-input cards let you provide a password, payment detail, or other sensitive value without intentionally putting it into ordinary chat. Fields marked sensitive are encrypted for delivery to the requesting execution host and represented to the agent by a handle. Remembered values are saved on the relevant computer using its supported protected credential storage. The agent can use the handle to enter the value into a designated website, command, or credential store. The destination then receives the value. Ordinary fields and any note you add may be visible to the agent. These features do not prevent disclosure if you paste a secret into chat, a tool reads an unrecognized secret from another source, or a destination exposes it in returned content.
5. Other recipients and service providers
In addition to the AI recipients in section 3, information reaches the following services when needed for the feature:
• Cloudflare: hosting, networking, databases and object storage, sandboxes, hosted sites, and related security infrastructure. • Stripe: hosted subscription payments, billing management, and payment-related records. • Expo and Apple/Google notification services: mobile push-token registration and notification delivery. Notification content and metadata may be processed by this delivery infrastructure according to the feature and your device settings. • Composio: authorization, connection records, action inputs, and results for integrations routed through Composio. First-party connector availability varies; not every connector bypasses Composio. • Exa and other configured search services: search queries, retrieved material, and request metadata for web search. • Slack and GitHub: Crew messages, workspace and repository authorization and content, commits or other requested account actions, and collaboration data. • Resend: email addresses and message content needed for sign-in, security, and other service emails. • Google: Google sign-in and connected APIs, configured AI services, maps features, and the limited website advertising measurement described in section 9. These are separate purposes and data flows. • Other services you choose: websites visited, apps and accounts connected, external MCP servers, recipients of messages or files, and providers you configure receive the information needed for the action you direct.
Third parties process data under their applicable terms and privacy policies. A connected service, recipient, or organization is not necessarily our subprocessor or under our control merely because Stella communicates with it.
We may disclose information when reasonably necessary to comply with law or valid legal process, protect safety and security, investigate abuse, enforce our agreements, or defend legal claims, subject to the stricter Google API rules below. If the business is involved in a merger, acquisition, or similar transaction, information may transfer subject to applicable law and privacy protections; Google API data remains subject to the Google API Services User Data Policy and Limited Use requirements.
6. Google Workspace connector and Google API Services user data
The following Google-specific disclosures and restrictions apply in addition to the rest of this policy. Our commitments on user-directed use, data processing, no generalized-model training, no sale or advertising or credit use, restricted human access, and Limited Use apply now to Google API data we receive, including through provider-routed integrations.
Google integrations currently use supported integration-provider routes, including Composio. Their authorization screens identify the permissions requested, and their applicable policies describe their credential processing. Connecting is optional. Disconnecting or disabling a service in Stella is not necessarily revocation of every scope in a shared Google grant; to revoke the grant itself, use your Google Account controls described below. Relevant content may still be retained in your conversation history after disconnection.
First-party connector disclosures for pending verification. The first-party Google Workspace connection flow is not currently enabled in the official hosted desktop application's connection wiring. Google OAuth verification and security assessment are in progress, not completed certifications. The first-party scope, granular-consent, and credential-storage disclosures below describe the connector covered by that pending verification and its intended supported behavior when released; they are not statements that every path described is available today or descriptions of the current Composio route. You do not need to enable this connector to use Stella.
The first-party Google Workspace connector covered by our verification is intended to let you optionally enable Stella to work with your Google account across Gmail, Google Calendar, Google Drive, Google Docs, Google Sheets, and Google Tasks. The following disclosures describe how that connector is intended to handle data received from Google APIs and apply in addition to the rest of this policy.
How You Connect — The connector is off until you choose to connect it, and connecting is always optional. Connecting starts a Google OAuth consent flow in which you sign in to Google and approve the access. You can connect Google services granularly — enabling one service at a time (for example, just Gmail or just Calendar), which incrementally expands a single shared Google grant to add each service's scopes — or use the one-tap Google Workspace bundle to enable all of the services listed above at once. A single consent screen does not necessarily cover every service; a granular connection requests only the scopes for the service you are enabling, and you review and approve the requested scopes on Google's consent screen before access is granted.
Data and Scopes We Access — When connected, only for the services you have connected and only as needed to carry out tasks you ask Stella to perform, the connector can access:
• Basic account identity — your Google account email address, basic profile, and an OpenID identifier, used to identify the connected account (scopes: openid, userinfo.email, userinfo.profile) • Gmail — read, compose, send, and organize your messages, drafts, and labels; this does not include permanently deleting mail (scope: gmail.modify) • Google Calendar — view and manage your calendars and events (scope: calendar) • Google Drive — view and manage files in your Drive (scope: drive) • Google Docs — create, read, and edit your documents (scope: documents) • Google Sheets — create, read, and edit your spreadsheets (scope: spreadsheets) • Google Tasks — view and manage your task lists (scope: tasks)
User-Directed Use — Stella accesses and acts on your Google data only after you connect the account and direct Stella (or a Stella agent acting on your instruction) to perform a task, such as reading an email, scheduling an event, or editing a document. Stella does not access your Google data for purposes you have not requested.
Where Your Credentials Live — Depending on how you run the connector, your Google OAuth tokens are stored in one of two ways. For on-device (local) execution, the OAuth access and refresh tokens Google issues are stored on your own device in Stella's protected credential store, encrypted at rest using your operating system's secure storage (the OS keychain or credential vault, via the desktop framework's safeStorage). For cloud or Store-based execution — where Stella carries out connector actions on your behalf from our backend — the tokens are persisted in FROMYOU's server-side credential vault, encrypted at rest using AES-256-GCM with a versioned master key. In both cases the confidential OAuth client secret is held only by FROMYOU's backend and is never shipped inside the app.
How Google Data Is Processed — Stella uses your connected Google account to carry out the tasks you request. Relevant content returned by Google APIs, such as emails, events, documents, or spreadsheets, may be processed by our services and stored in conversation history, tool results, or task records. When you direct Stella to reason over this content, it may be sent to the AI model provider handling your request and to service providers needed to operate that feature. We do not use Google Workspace data to develop, train, or improve generalized artificial-intelligence or machine-learning models.
No Sale, Advertising, or Credit Use — We do not sell or rent data received from Google APIs, and we do not transfer or use it to serve advertising or to determine creditworthiness or for lending purposes.
Restricted Human Access — FROMYOU personnel do not read data obtained through the Google connector, except where you give explicit consent to view specific data (for example, for support you request), where necessary for security purposes such as investigating abuse, where required to comply with applicable law, or where the data has been aggregated and anonymized for internal operations.
Retention, Deletion, and Revocation — Google tokens are retained while the connection remains active. Disconnecting the service removes the associated credentials. Google content included in conversation history or task records follows the retention and deletion practices described below; disconnecting a service does not delete that history. You can revoke Stella's access at any time from your Google Account under Security → Third-party access (https://myaccount.google.com/permissions), which invalidates the tokens held both on your device and in the vault; you can also disconnect individual services within Stella so Stella stops using those services. Disabling an individual service is not necessarily revocation of its OAuth scopes in a shared Google grant; use your Google Account controls to revoke the grant itself.
Limited Use — Stella's use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy (https://developers.google.com/terms/api-services-user-data-policy), including the Limited Use requirements.
If another part of this policy could be read to permit a use or transfer of Google API data inconsistent with these commitments, this section controls. General statements about service improvement, security, third parties, or business transfers do not authorize prohibited Google API data use or transfer.
7. Sharing and collaboration visibility
A public Drive share link publishes a separate copy that anyone with the link may access without an account. Deleting or changing the original Drive file does not necessarily remove that published copy. Hosted apps and previews, including on stellasites.app, may be public or audience-restricted depending on their settings. Do not assume a preview is private because its repository is private. People with access may download, copy, or redistribute content; removing a link cannot retrieve their copies.
Crew runs on the installing workspace owner's Stella account, using the model and engine selected for Crew and billing usage to that account. Each Slack workspace has separate Drive storage, notes and memory, working files, and cloud containers. Crew can use that workspace's files and rules, its connected GitHub repositories and configured repository environments, authorized search/maps/media/canvas tools, and its own published previews. A thread uses one connected repository and sees only its own agents. Replies and artifacts may be visible to Slack participants and the configured preview audience; repository actions appear in the connected repository.
Server-side access controls prevent Crew from accessing the owner's personal Drive, Stella memory, personality, skills, projects, apps or conversations; personal connected apps, MCP servers, saved secrets, browser/sign-ins, computers/devices or schedules; and another Slack workspace's storage or working files. This does not prevent the owner or a participant from deliberately supplying information to Crew. Older Crew files previously stored in the personal Drive's Slack workspace area are migrated into the workspace's separate storage, and the old synced Drive entries are removed.
Deleting your Stella account does not necessarily erase another participant's records, organizational Slack messages, repository history, public-recipient copies, or connected-service data. Contact us for help with content still hosted by FROMYOU; contact the relevant organization or third party for its copies.
8. How we use information and legal bases
We use the information described above to authenticate users; provide AI and agent features; synchronize and recover work; store and deliver files, messages, and apps; perform connected-service actions you direct; run schedules; process subscriptions; communicate service and security information; diagnose failures; prevent fraud and abuse; and meet legal obligations. We can use operational and telemetry metadata to understand reliability and improve the Service. This does not permit training generalized models on conversation content or prohibited use of Google API data.
Where EEA, UK, or similar data-protection law applies, FROMYOU relies as appropriate on:
• Contractual necessity to supply the service you request, maintain your account, and process billing. • Legitimate interests in security, fraud prevention, service reliability, and communicating with you, balanced against your rights. • Consent for optional data access or sharing where required, such as connected-account permissions and applicable AI-sharing, discovery, or advertising-measurement choices. You can withdraw consent for future processing; withdrawal does not affect lawful prior processing. • Legal obligations for required financial records, responses to legal process, and other compliance duties.
Some information is needed to provide the Service; declining required account information or task context can make a feature unavailable. Permissions you grant Stella do not eliminate any responsibilities you or your organization have for other people's data. If you use Crew for an organization, your organization may have separate controller responsibilities; this public policy is not a negotiated data-processing agreement.
9. Telemetry, browser storage, and advertising measurement
Optional telemetry records metadata such as event types, timing, tokens and cost, model or agent type, and error codes, with account or installation identifiers pseudonymized by hashing. Its defined event fields do not include chat messages, prompts, file content, tool arguments, or tool outputs. Hashing is pseudonymization, not a guarantee of anonymity. Settings → Privacy lets you turn off telemetry, remove queued device telemetry, and stop applicable future device and account-level cloud telemetry reporting. In-flight events may still arrive. Necessary billing, usage, security, and operational records still continue as described above.
Our apps and websites use cookies or local storage for sign-in, sessions, preferences, and necessary operation. Google Ads measurement is separate from assistant content and Google connected-account data.
The production website configures a Google Ads tag for advertising measurement. The tag can load for visitors generally, not only people who arrive through an ad. Google may receive page and device information ordinarily sent by your browser, identifiers or cookies associated with the tag, ad-click identifiers when present (such as GCLID, GBRAID, or WBRAID), and download or signup conversion events. Stella's signup-conversion deduplication flag in browser local storage has no application-set expiry and remains until cleared.
We use this measurement for advertising measurement and campaign optimization, not to serve advertising based on your conversations or Google connected-account data. We do not use enhanced conversions or intentionally send customer-provided email addresses to Google for this measurement. We do not configure this measurement for remarketing or targeted advertising. Google processes received information under its own privacy policy and terms. You can limit measurement by blocking advertising scripts or cookies and clearing this site's local storage and cookies. Where applicable law requires consent for the tag or related storage, that legal requirement applies; this policy itself is not consent.
10. Security and access
We use reasonable technical and organizational measures appropriate to the data and feature. These include TLS for hosted communications, encryption in server-side credential vaults, access controls tied to accounts and devices, device identity signatures, rate limiting, and application-level encryption of conversation records. Conversation encryption uses conversation-specific keys protected by our key system; deletion of a conversation removes its active key material. This does not mean every stored metadata record, media job, file, or public share has the same encryption scheme, or that the Service is end-to-end encrypted against FROMYOU.
Stella attempts to redact known secrets from model-visible or stored outputs, but redaction cannot guarantee recognition of every secret. Protect credentials and do not put them in ordinary chat or public files.
Current admin and support interfaces expose account, billing, and enforcement metadata rather than general conversation, memory, or file-content browsing. The privacy settings include time-limited support grants, with a 24-hour expiry, and a visible recent support-access audit log. A grant is not routine staff access to your content. Google-data human access remains limited to the exceptions in section 6.
No system is completely secure. Public content, third-party destinations, modified clients, and self-hosted configurations have their own risks. We do not promise that encryption, a sandbox, or an approval prompt prevents every unauthorized disclosure or harmful action.
11. Retention and deletion
Retention depends on the data, feature, purpose, available controls, legal requirements, and third-party policies. The hosted in-app account-deletion process has the specific recovery and processing timeline below; it does not promise immediate erasure from every storage layer or deletion of copies outside our control.
• Conversations, memory, task history, Drive, and hosted workspaces: kept to provide history, synchronization, ongoing tasks, and recovery until removed through applicable controls or account deletion, subject to legitimate legal, security, and operational needs. • Engine session records and checkpoints: can persist separately from a running sandbox or visible cache to enable continuation. Stopping a task or destroying a sandbox is not necessarily deletion of the associated history or workspace. • Connected-account credentials: kept while needed for the active connection. Disconnect removes the relevant active credentials according to the connection's controls; historical tool results and connected-service copies are separate. • Media inputs, outputs, and job records: kept to submit, deliver, manage, and troubleshoot the feature. Completed jobs and generated files can persist until removed; there is no blanket promise of automatic short-term deletion after completion. • Public shares, apps, and previews: may be separate copies from source files or projects. Remove the publication separately where controls are available, or contact us for assistance. • Account and authentication records: kept while your account is active and during the seven-day deletion recovery period. Sessions are revoked when the deletion request is recorded. Sign-in records are removed during final deletion, apart from the limited deletion, billing, and fraud-prevention records described below. • Billing, usage, and security records: kept as necessary for accounting, legal obligations, reconciliation, abuse prevention, and dispute resolution. A usage-limit reset is not deletion of historical billing records. • Temporary delivery and operational records: expire or are cleaned up according to the record type; some routing, deduplication, or recovery records persist longer than a stream or session. • Telemetry and support logs: metadata retained for their stated purposes. Turning telemetry off stops future optional reporting and removes queued device events; it is not a promise to erase all previously received telemetry or required operational records. The in-app support audit log shows a bounded recent history, not an unlimited archive. • Google Ads signup-deduplication flag: remains in browser local storage until cleared; Google-held data follows Google's policies. • Device-local data: remains until removed by you, an applicable client control, or your operating system; deleting a hosted account does not automatically erase all files on your devices. • Organizational and third-party records: subject to the organization or provider's controls. Crew uninstall or individual account deletion is not a guarantee that all workspace messages, installations, preview records, or repository history are erased.
Hosted in-app account-deletion timeline
When you request deletion in Settings → Account → Delete account, the server records the request first. If it cannot record it, the app shows an error with Retry and you remain signed in. Once recorded, your account is disabled immediately, all devices are signed out, agents, schedules, and syncing are stopped, hosted Stripe subscription renewal is turned off, and we send a confirmation email. If the initial Stripe renewal update fails, it is retried hourly during the recovery period.
You have seven days to recover the account. Signing in during that period cancels deletion, re-enables the account, and triggers a cancellation email. Subscription renewal stays off; you can manage or resume it in billing.
After the seven-day period, an hourly background job processes deletion, survives service restarts, and retries failures. It removes eligible hosted account content, including conversations, memory, agents and tasks, schedules, Drive and R2 files, hosted workspaces, connector and provider credentials, devices and pairings, owner-store content, and sign-in records such as sessions, linked logins, and passkeys. It also deletes the Stripe customer record, which cancels any remaining subscription. Some storage cleanup finishes asynchronously and is retried.
We send a completion email when that process is complete, usually within a day of the recovery period ending. This is the usual processing time, not a guarantee that every failure, third-party copy, or delivery issue is resolved within a day. If an expected email does not arrive or you need the status of a request, contact contact@fromyou.ai. Applicable statutory response and deletion requirements still apply.
After completion, we retain a minimal deletion record containing the account identifier and request/processing/completion timestamps, a minimal billing-account record and accounting information retained as required, and hashed abuse or fraud-prevention signals. The deletion record's email address is cleared after completion-email processing; delivery failures are retried before attempts end. Stripe retains its required billing/accounting records under its own policies and applicable law. These limited records do not include retained conversation, memory, or Drive content.
Email-only requests to contact@fromyou.ai are handled manually after any necessary identity verification. Deletion cannot recall information already sent to third parties or copied by recipients; local files, organizational records, and separately held third-party copies follow the applicable controls and policies. Store subscriptions, if offered, must also be managed with the store. Contact us if content you expect to be removed is still hosted by FROMYOU.
12. Your controls and privacy rights
• Delete your hosted account: use Settings → Account → Delete account in the official apps to start the seven-day recovery and background-deletion process described above. Instructions and an email request link are at https://stella.sh/delete-account. Email requests are handled manually; you do not need to reinstall a removed app. • Export and manage data: Settings → Privacy provides a JSON export of supported account data such as conversations, memory, and preferences. This is not necessarily a complete archive of every hosted file or third-party record. Download wanted Drive files separately and contact us for additional access or export requests. • Manage sessions: account settings let you review active sessions and sign devices out. Security emails may notify you of new sign-ins. • Control connected accounts and engines: disconnect services in Stella and revoke their authorization at the provider; change engines or remove your keys where supported. Google-specific controls are in section 6. • Control telemetry, discovery, and notifications: use privacy and onboarding choices and your device's notification settings. Notifications can expose information on a lock screen; choose settings appropriate to your use.
Depending on where you live and which laws apply, you may have rights to access or know the personal information we hold, obtain a copy or portability, correct it, delete it, restrict processing, object to processing, withdraw consent, or appeal a decision about your request. We will handle eligible requests as required by law, verify identity as appropriate, and explain any lawful exception. An authorized agent may make a request where permitted. We will not discriminate against you for exercising applicable privacy rights.
California residents may have rights under the CCPA/CPRA where applicable, including rights concerning sale or sharing for cross-context behavioral advertising and certain uses of sensitive information. We do not sell personal information or use your conversations for targeted advertising; our limited advertising measurement is described in section 9. We do not use or disclose sensitive information to infer characteristics about you for advertising. Contact us about applicable requests, including correction, deletion, access, or opt-out rights.
If EEA or UK law applies, you may also object to processing based on legitimate interests and complain to your local data-protection supervisory authority. Contact us first if you would like us to help resolve a concern. Withdrawing optional consent does not affect processing justified by another applicable legal basis.
13. Children's privacy
Stella is not directed to children under 13. We do not knowingly collect personal information from children under 13. Users must also meet the age and guardian requirements in our Terms and applicable local law. If you believe a child has provided personal information in violation of these requirements, contact us so we can investigate and delete it as appropriate.
14. International processing
FROMYOU is based in the United States. Cloudflare and other providers operate infrastructure in multiple locations, and information may be processed in the United States and other countries where we or the applicable provider operate. Processing location can depend on the service and configuration; we do not promise that all hosted data stays in a particular country. Applicable law may require safeguards for international transfers. Contact us for information about safeguards applicable to your data.
15. Changes to this policy
We may update this policy and will identify the revision date. We will provide reasonable notice of material changes through the Service or by email and obtain consent where legally required. A policy update is not retroactive consent to a new use of your personal information. Google API data remains subject to section 6 and the Google API Services User Data Policy, including Limited Use.
16. Contact
For privacy questions, rights requests, deletion help, or security concerns:
FROMYOU, LLC 131 Continental Drive, Suite 305 Newark, DE 19713, United States contact@fromyou.ai