Draft templates pending review by legal counsel — not legal advice. The sub-processor list is current and accurate.

Sub-processors

Third parties that process tenant data on WeShield's behalf. This list is current and accurate — data residency is Canadian except where noted.

Supabase

Canada — ca-central-1 (Montréal)

Primary Postgres database, authentication, and file storage

Data processed: All tenant application data, account identities, uploaded documents

Vercel

Canada — yul1 (Montréal)

Application hosting and serverless compute

Data processed: Request metadata; tenant data in transit during request processing

AI inference provider (configurable)

DeepSeek — People’s Republic of China jurisdiction. Anthropic — United States. AI inference is the one processing step that may leave Canada; it is opt-in per workspace and never receives un-redacted high-risk identifiers.

AI model inference for discovery, fit scoring, drafting, extraction, and the assistant. The active provider is a deployment configuration, disclosed here and in the AI-use registry. The current production provider is DeepSeek; Anthropic is the supported alternative.

Data processed: Prompts derived from tenant data, with high-risk identifiers redacted before transmission. Engaged only when AI features are enabled for a workspace.

Sentry

United States (or EU region) per configuration

Operational error monitoring (crash and exception telemetry)

Data processed: Application error events and stack traces. Personal information is scrubbed before transmission (no PII, no session replay); no tenant application data is sent.