FAMILYSIMPL
Your privacy
Last updated 17 November 2026. FamilySimpl is operated by Simpl Industries Pty Ltd, trading as Simpl Industries, in Western Australia, Australia. Contact contact@simpl.industries for privacy, access, correction, export or deletion requests.
Information collected and purposes
Account data includes your email, display name, account ID, verification and two-factor status, sign-in sessions and integration authorizations. Firebase Authentication handles sign-in credentials. We use these records to authenticate you, enforce permissions, send requested security messages and protect accounts.
You and authorized collaborators provide tree names, people’s names and aliases, dates, places, occupations, relationships, identity and link confidence, source URLs and archival locators, quotations, evidence, notes, research results and coverage, research reservations, explicitly shared AI summaries, claim reviews and conflict resolutions, original imports, portraits, logos and edit history. We use these to store and display your tree, assess recorded connections, organize shared research and apply requested edits. Connection colours and evidence-strength scores describe recorded evidence; they are not independent verification or a statistical probability of kinship.
Invitations include recipient emails, roles, capability links, expiry and use limits, delivery and acceptance records. Access requests and membership decisions are retained to enforce sharing. Requested custom hostnames include ownership challenges, DNS requirements and certificate/routing state. A hostname never makes a private tree public.
Service records include account and connection IDs, hashed session IDs, action/tool names, time, status, error codes, duration, model/token counts and estimated service costs. Assistant abuse controls use a keyed daily network identifier. Our usage ledger omits raw IPs and conversation bodies; infrastructure providers still process network and device information to deliver the service.
Optional Google Analytics receives fixed page/action labels and standard browser, cookie, device and approximate-location information. FamilySimpl sanitizes page paths and excludes family records, typed searches, emails, form content, private IDs and invitation tokens from intentional analytics parameters. Analytics is off until allowed on this browser and respects Do Not Track and Global Privacy Control. Analytics data may be exported to Google BigQuery for private reports.
Support email includes sender, recipient, subject, message, attachments and delivery data. Our Cloudflare email service forwards original messages, including attachments, to the Simpl Industries support mailbox. Its private database capture can contain a bounded plain-text body and attachment counts rather than attachment files. Forwarding and database retention are separate.
Living people, authority and collection notices
We receive family information from account holders, collaborators and imports, so a record may concern someone who has never used FamilySimpl. We use it for the family-history purposes above, keep access role-limited, and provide this notice to account holders and to affected people through our privacy request process. Uploaders should explain intended sharing to living relatives where appropriate and provide this notice. We assess additional direct notice when required and practicable; family research is not a blanket exemption.
Where consent is required, it must come from a person with capacity or lawful authority and be specific to the use. An uploader accepting our terms does not give consent for every relative. Sensitive living-person material needs the required documented permission or another applicable legal basis; collect and share only what is necessary. Ordinary names and dates are not automatically sensitive information, but can still identify living people. Records about deceased people can also reveal information about living relatives. Contact support if material is inaccurate, unexpectedly shared or lacks an appropriate basis.
We process essential account and service information to provide requested services, secure the service, administer rights and comply with applicable law. If EU/UK data-protection law applies to a particular processing activity, relevant bases can include contract for the account service, legal obligations, and legitimate interests in secure private family-history collaboration where a documented balancing assessment supports them. Optional analytics uses your browser choice. Sensitive information requires an additional applicable condition; we do not assume that an ordinary lawful basis or a genealogy purpose supplies it. We assess the activity and rights involved and can restrict processing while an issue is reviewed.
AI inputs, outputs and recipients
A connected ChatGPT or Codex host sends the arguments needed for each tool: selected tree/record IDs, queries, filters, requested edits, evidence or research text you ask it to save, and retry/revision identifiers. FamilySimpl does not require your full host conversation as a tool argument. The host receives privacy-filtered structured results, record/evidence IDs, uncertainty, research-status metadata, explicitly shared curated research/source summaries, minimal write receipts, and authenticated tree links. The interactive MCP Apps workspace displays bounded parts of the same filtered graph, shared evidence, research and reviewed change previews. Every server request uses the current account permissions; opening a view does not grant independent access to your account. It loads no remote family media.
Requested diagrams use eligible names, complete recorded birth and death dates, typed relationships, uncertainty and selected research or confidence metadata. Diagram requests can include a chosen direction, depth, reference person, named highlight groups, colours and matching rules. The host receives prepared diagram data and SVG panels and, where supported, a generated image or inline interactive view. These presentation requests do not change family records. A highlight shows a selection or recorded assessment, not a newly verified fact; exported or conversation-retained diagrams follow the destination’s retention and controls.
Workspace selections can send the selected eligible names and IDs, view state and a small generated context thumbnail to the AI host. Person-choice forms send the offered candidates, question suggestions and explicitly shared source summaries to the host; submitted answers return the selected IDs, question, research area and source identifiers to FamilySimpl. Source previews retrieve the same permitted shared summary. Preparing a packet or cancelling a form does not save family-record edits. Your question can subsequently be sent to the host when you choose a chat or draft action. The source picker does not upload files.
Presentation preferences, such as the default view, density, click behavior, date format and whether uncertain links are shown, are stored against your account in FamilySimpl’s database until replaced or removed with the account. Change them in workspace settings. The host can also retain view state under its own controls. Research files you explicitly open are parsed in the workspace; saving an opened host file or exporting a copy is a separate user action. Files and exported diagrams may contain the selected shared family information and follow the host or destination’s retention. Optional host-assisted summarization sends the request and selected shared context only after the host’s approval flow; its output is a draft and does not itself edit family records.
AI reads withhold under-13 and uncertain-age profiles and their links, private original narrative notes and citation text/URLs, quotations, original imports, portraits/photos, member details, change-history bodies and proposal-review content. Eligibility uses recorded dates, documented historical evidence, website age confirmation or an eligible recorded ancestry path. Parenthood alone does not establish an age. Incorrect records can affect this filter. Counts, paths and diagrams can cover less than the full website tree. A verified identity alone does not prove connection to you.
In Manage tree → Shared research workspace, authorized editors can deliberately share a separate curated summary of a saved research result, note or source. It can contain the research scope, finding, repositories/search terms and coverage, outcome, next step and an optional public citation URL or archival locator. Authorized AI connections and the website assistant can receive that summary to continue research; existing original notes are not automatically shared. Contributors propose these changes for admin review. Sharing does not change confidence or complete research. Known withheld identities and current record eligibility are checked, but lexical checks cannot recognize every sensitive statement: include only information intended for AI access.
Shared summaries remain with the source or research record until replaced or revoked. Revoke AI summary stops future retrieval of that summary; prior versions remain in the tree’s protected audit history, and copies already returned to a host conversation follow that host’s retention and controls. Membership or integration revocation stops future authorized access, not those earlier copies. General privacy/export/deletion requests remain available through support.
The read filter cannot prevent a person from typing sensitive material into an AI conversation or save request. Such text is processed by the AI host and relevant tool arguments reach FamilySimpl. Keep children’s private details, government identifiers, payment data, protected health information, account secrets and unnecessary living-person details out of AI requests. AI access is for genealogy, not DNA interpretation, sensitive-trait inference, medical decisions or tracking current locations.
The website assistant sends your question, bounded eligible tree facts, relevant recent context and proposed changes to OpenAI. Generation passes through Cloudflare AI Gateway; token counting also uses OpenAI. The code requests non-stored Responses API output and disables gateway payload logging and caching. Those settings do not promise zero provider retention. OpenAI’s default API abuse-monitoring retention can be up to 30 days, with exceptions; actual account controls govern. FamilySimpl does not train its own model on tree data. OpenAI API training/data sharing is governed by the provider account settings. Connected ChatGPT/Codex conversations follow your separate host settings and policies.
Who can receive information
Tree members receive content according to their role; admins and editors can apply permitted edits, contributors propose changes and viewers read. Invitation recipients/link holders receive invitation information. Simpl Industries support and authorized administrators may access information needed to handle support, privacy, abuse and reliability work.
Cloudflare supplies application delivery, D1 database, R2 object storage, queues, security, email, domain services and AI Gateway. Google/Firebase supplies authentication and website services; Google Analytics and BigQuery process optional analytics. OpenAI processes website assistant input/output, and an AI host you connect receives tool results. Email delivery/forwarding providers and our support mailbox process messages you send. Providers may process data internationally under their own contractual and legal requirements. We may disclose information when legally required or necessary to protect rights and security. We do not sell family-tree data or use it for targeted advertising.
Overseas processing
Likely overseas processing includes the United States: Firebase Authentication operates from US data centres, and our Google and OpenAI services may process account, service and assistant data there. Cloudflare delivers the service through its international network; routing, security and support processing can occur in Australia and other countries where its network, staff or subprocessors operate. Google website/analytics infrastructure and OpenAI subprocessors can also operate in other countries. We have not restricted all FamilySimpl processing to Australia and do not promise Australian-only or US-only storage.
Locations vary by service, selected region, routing and provider support. The current provider country/recipient details are available in Firebase privacy and security, Cloudflare’s processing terms and OpenAI’s subprocessor list. Contact us for the services and known countries relevant to your request. We assess appropriate contractual safeguards and applicable transfer requirements; this disclosure is not a waiver of our responsibilities or your rights. A connected host you choose has its own arrangements and controls.
Retention
Active family records, sources, imports, images, research, memberships, edit history and operation receipts remain while the account/tree needs them, or until a verified deletion request is resolved. Archiving, exclusion, resolving a note, disconnecting an integration or leaving a tree does not erase shared evidence. Operation and applied-assistant receipts are retained with the tree to prevent duplicate edits. Unresolved security/provider transactions can be retained until reconciled.
Website assistant context expires from reuse after one hour; daily bounded housekeeping removes expired context. Raw service-usage events are eligible for removal after 30 days; account-linked hourly aggregates after 365 days. Completed/failed/incomplete assistant request ledgers are eligible after 90 days; unresolved reservations remain for reconciliation. Incoming database email records are eligible after 365 days and outgoing metadata after 90 days. Daily batches are capped at 5,000 rows per category; failures or a backlog can delay physical removal. They never delete family trees. Support-mailbox copies and provider backups/logs follow their separate retention and legal requirements; the application cleanup does not erase those copies. We do not promise an unverified backup-erasure deadline.
Google Analytics, BigQuery, authentication, email and AI providers apply the retention configured for their services. Application row cleanup does not control these independent copies. Contact support for the current provider-specific retention relevant to a request. Disconnecting an AI host stops future authorized calls, not results already in that host’s conversations.
Device storage and controls
The website can keep loaded owned-tree records, reduced images, saved views, drafts and pending edits in account-scoped IndexedDB for offline use. Shared trees remain online-only. The public app shell may be cached. Signing out stops access in the application but can preserve the original account’s pending work for later sync. Local storage is not protection against someone controlling your unlocked browser profile. Clearing site data removes local drafts and pending edits: sync pending edits or copy drafts first.
Correct records through the tree’s editing controls and revoke invitations or memberships through sharing controls. Tree settings → Details can download current saved tree records as JSON, including excluded and archived records; media file contents, revision history, account information, access credentials and unsent device edits are excluded. Account-wide export/deletion requests are handled through support after identity and shared-record checks; there is no automatic account-wide delete button. We explain any information retained for shared rights, security or legal reasons. New account holders must be at least 18 and have reached the legal age of majority where they live. Existing account holders receive notice and an opportunity to review the updated account terms; ask support about export or transfer if you cannot accept them. Tell support about a suspected child-data or privacy error.
Your rights and privacy complaints
You do not need an account to raise a concern about a record concerning you or your dependent. Email contact@simpl.industries with “FamilySimpl privacy request” or “FamilySimpl privacy complaint”, a safe way to contact you, the relevant tree/record if known, and the outcome you seek. Do not initially send passwords, identity documents or private family evidence. We request only proportionate identity or authority checks before releasing or changing protected data.
We aim to acknowledge complaints within five business days and give a substantive response within 30 days, explaining any delay. The process is: confirm scope and identity as needed; investigate access, sharing and provider copies; consider affected people’s rights; explain the outcome, remedy, any lawful retention and review options. Tell us if urgent safety or child-privacy concerns require prompt restriction. An administrator cannot veto a valid privacy request merely because they own the tree.
Depending on applicable law, you may request access, correction, deletion, restriction, objection, portability or withdrawal of a consent-based use. Withdrawal affects future processing and does not itself erase lawful past processing, shared audit records or host copies. If we cannot meet a request, we explain the reason and available review route. You may complain to the Office of the Australian Information Commissioner where it has jurisdiction, or the competent regulator or court available under your local law. This does not promise that every regulator has jurisdiction over every case.
Agreement and permission records
We keep the account ID, document versions and hashes, time and flow of affirmative terms acceptance, adult-account declaration and privacy-notice acknowledgement. These are separate from optional analytics and AI read/write choices. We archive the exact legal documents and retain agreement evidence while the account is active and for as long as reasonably needed for legal obligations or a dispute, reviewed on closure. We do not collect government identity documents or raw IP addresses just to prove routine acceptance. Earlier accounts retain their historical acceptance version; we do not invent a new acceptance on their behalf.
We update this policy when features or processing change. Review the date and your choices before connecting an AI host. See OpenAI’s privacy policy, OpenAI API data controls, Cloudflare’s privacy policy and Google’s privacy policy.