Subprocessors & Data Processing Overview

Rakenne relies on a small set of specialist vendors to run the rakenne.app platform. This page describes which vendors we use, what role they play, and what types of personal data they can see when acting as our subprocessors.

This overview lives alongside (and does not replace) our Privacy Policy , which explains legal bases, retention rules, and your rights in more detail.

How We Use Subprocessors

We engage subprocessors only where they add clear value to the product (for example, secure hosting or payment processing), and we:

Limit each provider to the minimum data they need for their task.
Put written data protection terms in place (including DPAs and, where relevant, SCCs).
Review providers periodically as part of our vendor-risk process.

At a high level, subprocessors support five areas of the product:

Infrastructure & storage for the core application and workspace data.
AI inference for skills and chat-based workflows.
Transactional and inbox email for account-related communication.
Billing and subscriptions for paid plans.
Website and product analytics to understand how people discover and use Rakenne.

Infrastructure & Hosting

Google Cloud Platform (GCP)

Purpose: Application hosting, persistent storage, networking, backups, secrets management, DNS.
Services used: Compute Engine, persistent disks, Cloud Load Balancing, Cloud DNS, Secret Manager, networking components.
Categories of data: Account data (e.g. email, name), usage metadata, project configuration, document content, access logs, encrypted secrets.
Location: United States (GCP regions).
Retention: Application data retained according to our product data retention practices; infrastructure logs retained according to GCP defaults and our logging configuration.
Notes: Data is encrypted in transit (TLS) and at rest. Access to GCP resources is restricted via IAM and least-privilege principles.

AI Processing

Rakenne uses Google Gemini as the default large language model (LLM) for processing prompts, skill instructions, and document context. Google Cloud AI acts as a subprocessor for your content in order to generate responses.

Legal basis: Performance of contract (providing the core AI-assisted drafting functionality).
Categories of data: User prompts, skill metadata, document fragments, system instructions, and related metadata necessary to generate responses.

Google Cloud AI (Gemini)

Purpose: Default AI inference for skills, chat, and workflow steps — processing prompts and document context to generate drafts and responses.
Location: Google Cloud regions as configured by Google; typically United States or EU-based regions for model hosting.
Retention: Under Google’s enterprise AI terms, data is retained for limited periods for service operations and abuse prevention and is not used to train base models.
Notes: We supply a dedicated API key managed via Google Secret Manager.

Bring Your Own Key (BYOK)

Rakenne also supports Bring Your Own Key for Anthropic (Claude), OpenAI, and Google. When you configure your own API key, the LLM provider acts under your agreement and is not our subprocessor for that processing path. In that case:

You are responsible for configuring data retention, regions, and compliance options with that provider.
Rakenne simply forwards requests and responses between your workspace and your chosen provider.

Email & Communications

Mailgun

Purpose: Transactional email delivery (signup confirmations, email verification, and similar messages).
Categories of data: Recipient email address, message content, headers, and delivery metadata (timestamps, status codes, bounce information).
Location: United States and EU Mailgun infrastructure (routing depends on configuration and Mailgun’s network).
Retention: Email logs retained by Mailgun for a limited period under its standard terms (e.g. days to weeks) for delivery, troubleshooting, and abuse prevention.
Notes: Mailgun API keys are stored in Google Secret Manager and injected at runtime; emails are sent from mail.rakenne.app.

Private Email / Namecheap (Email Hosting)

Purpose: Hosting of mailbox-based email addresses (e.g. support or contact inboxes) used for customer communications and DMARC/SPF alignment for certain marketing domains.
Categories of data: Email content, headers, and metadata for messages exchanged with our support and contact addresses.
Location: Namecheap Private Email infrastructure (data centers operated by Namecheap and its subprocessors).
Retention: Emails retained in mailboxes according to our internal retention practices and legal obligations.
Notes: This infrastructure is separate from Mailgun, which is used for automated transactional emails from the app.

Billing & Subscription Management

Stripe

Purpose: Payment processing, subscription billing, invoicing, and billing-related notifications.
Categories of data: Customer identifiers, email address, subscription plan and status, partial payment method details (e.g. last 4 digits, card brand), invoices, and payment events. Full card details are handled exclusively by Stripe.
Location: Global infrastructure with EU data centers; Stripe acts as a GDPR-compliant processor with appropriate transfer mechanisms.
Retention: Billing-related data is retained for the duration of your subscription and for at least 7 years thereafter to comply with tax and accounting requirements.
Notes: Rakenne does not store full payment card numbers or CVV codes.

Analytics & Telemetry

Google Analytics & Google Tag Manager

Purpose: Measuring aggregated website traffic, campaign performance, and user journeys. Tag Manager is used to orchestrate analytics tags.
Categories of data: Pseudonymous identifiers, page URLs, referrers, device and browser information, approximate location (based on IP), and event data (page views, clicks). No document content or chat messages are sent.
Location: Global Google Analytics infrastructure, including servers in the United States and EU.
Retention: Aggregated analytics data retained according to our Google Analytics property settings (typically 14-26 months); Tag Manager configuration data retained while the container is active.
Notes: Where required, analytics is subject to cookie consent. IP addresses are processed according to Google Analytics configuration (e.g. IP anonymization).

Data Protection Practices

Minimised scope: Each provider receives only the data it needs to perform its service (for example, Stripe receives billing identifiers but not workspace documents).
Contractual safeguards: Our agreements include appropriate data protection terms, including Data Processing Agreements (DPAs) and, where required, Standard Contractual Clauses (SCCs) for international transfers.
Security review: We select vendors with strong security postures and reassess them periodically as part of our security and compliance work.
No LLM training on your content: When we use third-party LLM APIs with our keys, we select configurations and terms that state your prompts and documents are not used to train base models.

Keeping This List Up To Date

Our infrastructure evolves over time. When we add, replace, or remove a subprocessor in a way that materially changes how your data is handled, we will:

Update this page to reflect the new provider and role, and
Provide reasonable advance notice through in-app messaging or email where required by law or contract.

You can reach us at privacy@rakenne.app if you have questions about a specific provider or need a signed DPA that references this subprocessor list.