Loopclub Ltd View as Markdown
Subprocessors
Effective 11 October 2026.
Loopclub Ltd, 4023 Kennett Pike #50389, Wilmington, DE 19807, USA, uses the providers below for the purposes described. This page distinguishes customer-data subprocessors, our own infrastructure and other recipients. The DPA governs authorisation, objection and at least 30 days' advance notice to affected customers of new or replacement subprocessors. Questions: legal@doubleagent.so.
Customer-data subprocessors
| Provider | Purpose | Data involved | Location |
|---|---|---|---|
| Cloudflare, Inc. | API and delivery network; website/dashboard hosting; Workers, D1, KV and Queues; authentication email and email routing | Service traffic in transit; account/site settings, authentication and audit records, temporary presence, queued events and operational metadata according to the service used | Global network; headquartered in the USA |
Infrastructure operated by Loopclub Ltd
Loopclub Ltd self-hosts Elasticsearch in Canada for session and event records, edge requests, raw connection/device records, summaries, directory data and feedback. Self-hosting describes operation of the database; it does not establish that every upstream facility or infrastructure provider is outside the subprocessor chain. Any third party processing customer data for that infrastructure must be disclosed and covered appropriately before use.
The standard retention commitments are in the privacy policy. Backup retention is a separate operational schedule and must not be assumed to equal the live database periods. Contact legal@doubleagent.so for the applicable backup and infrastructure details.
Optional model-based processing
The reviewed service configuration disables the second-opinion judge. Supported backends include TypeSafe (Jev) and Cloudflare Workers AI; their presence in the code does not mean they are active or that a customer can enable an unapproved processor.
If enabled, a judge can receive page/referral metadata, browser-environment measurements, interaction statistics, numeric features, signal codes and a summarised timeline. The standard request does not deliberately include top-level raw IP, full user agent or linked order identifiers. Page and environment fields must still be reviewed for incidental personal data; there is no blanket guarantee that evidence is anonymous.
Before activation, we must confirm the provider's contracting entity, location, retention, training restrictions, security and transfer arrangements, update this list and complete the DPA's notification process. TypeSafe is not represented here as an approved active subprocessor with a verified processing location. Cloudflare Workers AI would extend Cloudflare's stated processing purpose and must be reviewed before activation too.
Public identity sources and lookup recipients
Identity and directory features retrieve public keys, agent cards, registry data, metadata and reviews. Current Ethereum mainnet and Sepolia RPC defaults use PublicNode; configured alternatives, metadata publishers and gateways may also receive lookups. Recipients see requested wallet/contract/token or resource identifiers and the requesting server's connection metadata. The ordinary lookup is not intended to send a visitor's session dossier.
These recipients operate external services and publish their own terms. They are not all automatically contracted subprocessors, and being a public endpoint is not a privacy exemption. Their role and transfer arrangements must be assessed for the particular data flow. Public blockchain information and third-party copies cannot be deleted by Loopclub Ltd.
Tools selected by customers
Google Analytics, Meta, Shopify, Stripe, Klaviyo, Mixpanel and other connected destinations receive the identifiers, labels or events specified by the chosen integration. Customers select and control these connections under their own arrangements; they are not automatically Loopclub Ltd's subprocessors merely because we support an integration. Google Search Console works the other way round: when a customer connects it, we read that customer's aggregated search performance data (queries, pages, countries, devices and click counts per day) from Google with a read-only OAuth grant the customer can revoke.
Analytics on Double Agent's own sites
Our production website and dashboard load Google Analytics 4 (Google) and Microsoft Clarity (Microsoft) for usage analytics and, for Clarity, interaction/session-replay analysis. They may receive information about account holders visiting the dashboard, not just anonymous marketing-site visitors. Selected app content and inputs are masked; masking does not exclude all customer-related information or eliminate network metadata.
These are recipients for our own website processing, distinct from a customer's installation. Consent, applicable opt-outs, provider settings, contractual terms and international-transfer safeguards still apply. See the privacy policy, Google's privacy policy and Microsoft's privacy statement.
Team notifications
We post account events to our own team's workspace on Slack (Slack Technologies, LLC), so we can help new customers and act quickly on billing and domain problems. These messages carry our customers' account contact details: account owners' email addresses, account names, domains, and plan, payment and domain events. They do not carry visitor sessions or any other data we process for a customer's site.
Slack is a recipient for running our own customer accounts. It is not a customer-data subprocessor, so the DPA's notice period does not apply to it. See Slack's privacy policy.