Privacy Policy
This policy explains how DeepLAN processes information. DeepLAN is a local-first macOS network inventory and diagnostics application. It does not require an account to scan an authorized network or use its sample mode.
Local network information
When you authorize a network, DeepLAN may process local network addresses, hardware addresses, device and service names, service metadata, bounded probe results, local HTTP titles or banners, reachability results, timestamps, and the labels, rooms, trust states, findings, and diagnostics you create. When you pair an optional local source, it may also process the source’s reported device role, model, firmware, serial or hardware identity, address, health, and coverage information.
This information is used to build the inventory, show evidence and changes, and run the diagnostics you choose. It is stored in DeepLAN’s sandboxed shared container on your Mac. Sample mode uses built-in fictitious data and does not scan a network.
DeepLAN does not capture general packet payloads or browsing content. It does not include advertising, cross-app tracking, or product-analytics SDKs. Its managed camera inventory paths do not request or retain video, audio, recordings, snapshots, streams, PTZ/talkback data, or camera settings.
Local retention and your choices
Inventory, observations, settings, and history remain on your Mac according to the retention choices available in DeepLAN. Revoking a network authorization or disconnecting a source stops future collection; it does not by itself erase existing local history. Exports are written only to the location you choose in the macOS save panel and remain there until you remove them.
You can keep DeepLAN in sample mode, decline Local Network access, revoke a network authorization, disconnect a paired source, and leave optional remote AI disabled. Credentials you supply for a provider or controller are kept in the macOS Keychain rather than in DeepLAN’s inventory database, logs, or sample data.
Optional Hosted Intelligence
Hosted Intelligence is optional and off by default. Before each request, DeepLAN presents the exact selected evidence records, provider, payload size and digest, and credit cost, then asks for explicit foreground approval. Depending on the operation, those records can include a selected device identifier and observed identity/service metadata; a recorded finding and its coverage limits; bounded current-device summaries for a question; event-ledger rows within a stated brief window; a correlated incident’s retained observation window; or identifiers, digests, state, and event history for a selected staged policy. It does not include general network traffic contents, credentials, controller commands, or the rest of your inventory.
With your approval, the bundle is sent over HTTPS to DeepLAN’s fixed service at
api.deeplan.app, which uses Anthropic’s commercial API solely for the
requested identification, explanation, query, brief, incident, or policy-analysis
operation. DeepLAN does not accept arbitrary provider, model, route, or tool
choices for Hosted Intelligence. A weekly or biweekly reminder never sends
automatically; each due brief still requires a new foreground disclosure and Send
decision. AI output cannot create a finding or authorize or enforce a network
change.
The service maintains a pseudonymous metering and anti-replay ledger. It does not retain raw StoreKit proofs in that ledger. Completed result bodies are swept no later than 30 days after provider dispatch; a pseudonymous terminal digest and anti-replay binding can remain so a charged request cannot run again.
As of August 30, 2026, Anthropic states that its commercial API deletes inputs and outputs from its backend within 30 days by default. Anthropic identifies exceptions for a separately agreed retention setting, services with longer customer-controlled retention, usage-policy enforcement, and legal obligations. Anthropic states that usage-policy-flagged inputs and outputs may be retained for up to two years and related trust-and-safety classification scores for up to seven years. Anthropic also states that commercial API inputs and outputs are not used to train its models by default unless the customer affirmatively allows that use or submits feedback. DeepLAN does not submit provider feedback or opt this API traffic into model training. Review Anthropic’s current API retention information and commercial data-use information.
A bring-your-own provider key, if you choose to configure one, can associate a direct request with your provider account. You can decline the per-request disclosure or select local-only operation to prevent future optional AI requests.
Sharing and security
DeepLAN does not sell personal information or use it for advertising or tracking. It discloses information only when you choose an optional provider or source, to service providers needed to operate a requested DeepLAN feature, or when required by law. DeepLAN uses the macOS App Sandbox, Hardened Runtime, Keychain, and operating-system security APIs to protect local information. No storage or transmission method is completely secure.
Children
DeepLAN is a general-audience network utility and is not directed to children.
Changes and contact
Material changes will be reflected by updating this effective date and, where appropriate, by notice in the app or on this website. Email edge@aerodyneai.com for privacy questions or deletion requests, or visit DeepLAN Support. Do not send credentials, private keys, raw StoreKit receipts, or an unredacted network export by email.
This policy describes DeepLAN’s current release configuration. If a feature, provider, retention period, or data flow changes, this policy will be updated before that change is released.