Stays local: credentials and source
AI provider credentials, API keys, GitHub repository contents, and website source code are not sent to letAGENTcook servers.
Security and privacy
What Stays Local and What letAGENTcook Services Process
letAGENTcook is a local-first desktop app. Your credentials, source code, analytics data, store credentials, prompts, raw AI answers, reports, and full tracking history stay on your devices; Usefo Inc. services process only the account, license, update, security, refund, and support data required to operate the product.
Working modelLocal desktop app
Service roleSign-in · licensing · updates · support
Future optimization dataNot enabled
Data boundary
Keep working data on your devices. Send only the service data the product needs.
Direct connections you authorize may send requests from your device to an AI provider, Google Analytics, GitHub, App Store Connect, or Google Play. Those third-party connections are separate from data sent to letAGENTcook servers.
Stays local: analytics and stores
GA4 raw data, App Store Connect credentials, Google Play credentials, and raw store-console data remain in the authorized local workflow.
Stays local: prompts and history
Prompt text, raw AI answers, full reports, full tracking history, and brand, domain, product, website, and app analysis data stay on your devices.
Service data: account sign-in
Apple, Google, or GitHub sign-in requires an email address or account identifier and the session data needed to provide account access.
Service data: licenses and security
Purchase, selected domain or app, activation, binding, validation, refund, chargeback, and abuse-prevention records support the permanent license contract.
Service data: updates and support
Usefo Inc. services deliver App, Agent, Skill, Capability, and Workflow updates and process the minimum evidence you choose to provide for support, privacy, security, refund, or correction requests.
Future data policy
De-identified Auto Research and GEPA data collection is not enabled.
No product-generated optimization data should be uploaded until the policy, implementation, purchase consent, and legal basis are complete and publicly documented.
- Publish a field-level data dictionary and explicitly excluded categories.
- Document the de-identification method, purpose, retention period, and security controls.
- Exclude credentials, source code, GA4 raw data, store-console raw data, prompts, and raw AI answers.
- Complete legal review and show clear purchase consent before any collection begins.
- Do not claim SOC 2, ISO 27001, an independent security audit, or another certification unless a public verification source is available.
# future optimization data status
$ collection status: not enabled
✓ data dictionary pending
✓ de-identification rules pending
✓ retention and legal basis pending
No upload until the public policy and implementation are complete.