Privacy
How UmmahSignal is designed to process workspace and configured public-source data.
Company and contact
UmmahSignal is operated by DAIDALOS LIMITED, company number 17495199, incorporated in England and Wales on 2 October 2026.
For questions about the service or privacy requests, contact support@ummahsignal.com.
What we process
For approved production workspaces, UmmahSignal processes account and organisation details, billing and entitlement state, monitor terms, workflow actions, alert recipients, API and webhook configuration, and configured public content from official X, YouTube, Reddit and LinkedIn interfaces when each source-specific access gate is satisfied.
Public-source records can include the public text, source URL, public author metadata, timestamp and engagement metadata supplied by the provider. We do not access private messages, protected content or private groups, and we do not build cross-platform author profiles.
Purpose and responsible boundaries
We use the minimum necessary public content to match configured monitors, remove duplicates, classify only policy-approved eligible content, support human review, deliver requested alerts and webhooks, and prepare workspace analytics.
UmmahSignal serves brands, apps, creators and organisations and must never infer, label or profile an individual’s religion, ethnicity or another sensitive trait. YouTube supports exact matching and manual review. LinkedIn is limited to separately authorised Page management. Neither source supplies AI enrichment, cross-source analytics, exports, REST/MCP data or external signal notifications. Reddit AI processing requires separate explicit policy approval; without it Reddit remains exact/manual review.
Retention and deletion
YouTube records and their linked workflow data expire after 29 days and are erased together, including source identifiers and URLs. LinkedIn Page activity expires after 47 hours; legacy LinkedIn records in the general inbox expire within 24 hours. Those deadlines continue to apply to read-only and canceled workspaces.
For other approved sources, raw public text, public author metadata, engagement data and embeddings are retained for no more than 30 days; workflow metadata and classification corrections may be kept for up to 12 months, subject to any shorter provider agreement. Content-free audit history may be retained for up to 12 months.
Workspace deletion immediately stops processing and revokes active credentials. Active workspace data is scheduled for deletion within seven days and backup copies expire within 30 days. Canceled workspaces remain read-only for up to 30 days unless the owner requests earlier deletion.
YouTube API Services
UmmahSignal uses YouTube API Services to retrieve public video details and public comments. We display these only within the authenticated product for human review, with links to the original YouTube content. We do not download video or audio, request private YouTube account data, or use this content for AI processing.
Google handles API requests under its privacy policy. The current public-data integration does not request Google OAuth consent. If you have granted access to another Google-connected application, you can manage that access in Google's security settings.
Facebook and Instagram connected accounts
When a workspace owner connects accounts through Facebook Login for Business, we process the selected Page and linked professional Instagram account identifiers, names, available account photos, granted permissions and encrypted access tokens. These credentials are used by our servers to retrieve authorised content and receive signed activity notifications; they are not shown to other workspace members.
The integration retrieves supported comments on selected Facebook Page posts, comments on connected Instagram media and supported Instagram @mentions. Records include provider identifiers, content text, source links, timestamps and author information supplied by Meta. Stories, private-account mentions, private messages and publishing are not supported. Initial and scheduled syncs are bounded; the coverage view explains when results may be incomplete.
This content supports the private workspace inbox, human assignment and review, account-filtered analytics and requested email alerts. Facebook and Instagram content is not sent for AI enrichment, exposed through REST/MCP, included in exports or published on public dashboards. Email alerts link back to the authenticated inbox without including the comment text.
Facebook and Instagram content and its associated review records expire after 29 days. An owner can disconnect in Settings → Connected accounts. This immediately stops collection and schedules erasure of stored content, identifiers and linked review records for the connection. Revoking the app through Meta or submitting Meta’s data deletion request also revokes access and starts erasure; the deletion callback returns a status link.
Website analytics
We use Zijra to measure visits to public website pages and beta-access actions in aggregate mode, without a persistent visitor identifier. Private workspace, inbox, administration, sign-in and invitation routes are excluded from this installation.
Page paths exclude query parameters and URL fragments. The tracker respects Do Not Track and Global Privacy Control signals. Our beta-application event contains only the selected beta or demo intent and the form placement; it does not include names, email addresses, application identifiers or form contents. We do not send user identification or payment information to this tracker.
For supported AI and search crawlers, our server sends the requested public page path, time, user agent and trusted source IP to Zijra for independent identity verification. Addresses and user agents are encrypted before entering our delivery queue and removed after delivery or within 23 hours. Zijra does not store the IP or raw user agent in the crawler report; it records only verified provider, purpose, public page and time. Private account routes, cookies, query strings and customer content are excluded.
AI and service providers
Eligible X content, and Reddit content only after separate policy approval, may be sent to OpenAI for schema-validated classification and summarisation with storage disabled at the API request level. Author identity is omitted wherever practical. Anthropic is an optional disabled fallback and cannot be activated without an equivalent privacy review.
Production subprocessors are expected to include Clerk, Convex, Vercel, Stripe, OpenAI, Resend and Sentry, alongside X, Google/YouTube, Reddit, LinkedIn and Meta as source providers when enabled. Their final roles, locations and data-protection terms must be confirmed in the launch subprocessor list.
International processing
Convex EU West is selected for regional proximity, but UmmahSignal does not promise that all processing remains in one country or region. Our providers may process data internationally under their contractual transfer mechanisms.
Your controls
Workspace owners can request exports, workspace deletion and review of a public-content URL. X compliance removals have a 24-hour internal target; YouTube, Reddit, LinkedIn and individual public-content requests have a seven-day target, subject to identity, source and legal checks.
Requests should be submitted through the in-product removal flow or the support channel named in the customer agreement. We may retain limited content-free records where required for security, billing, legal claims or proof that a request was completed.
Legal review before paid launch
This is a product-ready privacy draft, not legal advice. A qualified reviewer must approve the lawful bases, controller and processor roles, contact details, cookie position, subprocessor list, international-transfer language and jurisdiction-specific rights before paid production launch.