Service Providers & Subprocessors

Core platform and data hosting

• Lovable — Hosts and deploys the Jorvea application and provides the AI gateway used by the default Gemini model route. Data categories: Request metadata and, for AI features routed through the gateway, the job, resume, profile, answer, and draft content included in the prompt.
• Supabase — Provides the managed PostgreSQL database, authentication, session management, and file storage. Data categories: Account and authentication records, career profiles, resumes, opportunities, applications, outreach and inbox records, billing/usage records, and uploaded files.

AI model and research providers

The default model route uses the Lovable AI gateway and a Google Gemini model. When separately configured, Jorvea can call OpenAI, xAI, or Anthropic directly as a preferred or failover provider. Research workflows may call Perplexity separately. A provider is not used unless its route is configured and invoked.

• Google Gemini, through the Lovable AI gateway — Generates or tailors application materials, screening answers, outreach drafts, interview preparation, and in-product coaching. Data categories: Job posting text, resume and profile content, prior answers, and the material being generated.
• OpenAI direct API — Configured preferred or failover route for supported AI generation workflows. Data categories: The same prompt content required by the applicable AI feature. Conditional — only when a direct OpenAI API key is configured and the route is selected.
• Anthropic direct API — Configured preferred or failover route for supported AI generation workflows. Data categories: The same prompt content required by the applicable AI feature. Conditional — only when an Anthropic API key is configured and the route is selected.
• xAI direct API — Configured preferred or failover route for supported AI generation workflows. Data categories: The same prompt content required by the applicable AI feature. Conditional — only when an xAI API key is configured and the route is selected.
• Perplexity — Produces company or role research for workflows that enable external research. Data categories: Company and role identifiers and public job text included in the research query. Conditional — only when the research integration is configured and requested.

Payments

• Stripe — Handles subscription checkout, billing, billing-portal access, boost-pack purchases, and payment-status webhooks. Jorvea does not receive full card numbers. Data categories: Email address, customer/subscription identifiers, selected plan, purchase identifiers, and payment status. Stripe collects payment-card details directly.

Email and messaging

• Resend — Sends transactional messages and outreach emails that a user approves or authorizes. Data categories: Recipient address, subject, message body, reply-to data, and delivery status.
• Mailgun — Routes inbound mail for managed applicant inbox addresses, including employer replies, confirmations, and verification messages. Data categories: Sender and recipient addresses, headers, subject, message body, and provider message identifiers.
• Twilio — Sends SMS alerts or confirmations to users who separately opt in. Data categories: Phone number, SMS content, and delivery status. Conditional — only when SMS is enabled and the user has consented.

Connected accounts

• Google APIs — Supports user-authorized Gmail and calendar workflows, such as retrieving relevant verification messages or creating calendar events. Data categories: OAuth tokens for the scopes granted and the specific message or calendar content covered by those scopes. Conditional — only after the user connects a Google account.
• LinkedIn — Supports LinkedIn-related posting or outreach where the connected account and available API permissions allow it. Other workflows use a manual handoff that opens LinkedIn for the user to send. Data categories: OAuth tokens and content the user chooses to publish or send through an available API route. Conditional — only after the user connects LinkedIn and the requested API operation is permitted.

Opportunity discovery and enrichment

These integrations receive search criteria, public URLs, or company/role identifiers. They are not intended to receive a user's resume unless a specific workflow clearly says otherwise.

• Job-source APIs, including Adzuna, Jooble, Careerjet, and TheirStack — Retrieve job postings that match user search criteria. Data categories: Keywords, locations, work preferences, and other search filters. Conditional — only when the applicable source is configured and queried.
• Contact-enrichment providers, including Apollo, Hunter, ContactOut, RocketReach, Snov, and NinjaPear — Find business contact information for recruiter or hiring-manager outreach. Data categories: Company names, domains, and role titles used as lookup inputs. Conditional — the active provider depends on configuration and availability.
• Public-web retrieval providers, including Firecrawl, Tavily, and ScrapingBee — Retrieve public job or company pages for supported discovery and research workflows. Data categories: The public URLs or public search queries being retrieved. Conditional — only when the applicable retrieval route is configured and invoked.

Application automation and background workflows

• Hosted browser providers, including Browserless and Browserbase where configured — Run cloud browser sessions that open and complete supported application forms. Data categories: Application-form content, resume or cover-letter files, screening answers, and site session state or credentials supplied for that application flow. Conditional — only for cloud-automation workflows using the configured provider.
• CapSolver — Handles supported challenge prompts encountered by an automated application flow. Data categories: Challenge parameters and page identifiers required to process that challenge. Conditional — only when the feature is configured and a supported challenge is encountered.
• Inngest — Schedules and runs background workflows such as discovery, dispatch, and follow-up jobs. Data categories: Internal record identifiers and workflow parameters included in queued events.

Error monitoring and product analytics

These browser SDKs initialize only after analytics consent. Global Privacy Control is treated as a decline, PostHog autocapture and session recording are disabled, and withdrawal stops future capture.

• Sentry — Application error and performance monitoring. Data categories: Scrubbed error messages, stack traces, page path, and limited technical context. Conditional — consent-gated.
• PostHog — Curated product-usage analytics. Data categories: Explicit event names and scrubbed properties, page paths, and a device or user identifier. Conditional — consent-gated; DOM autocapture is disabled.