Data Subprocessors · TeamTeacher

TeamTeacher uses the following trusted service providers ("subprocessors") to help deliver our AI-powered teaching platform. These subprocessors handle specific aspects of our service and operate under strict data processing agreements that comply with our Privacy Policy and applicable data protection law, including GDPR.

Infrastructure & Data Processing

Supabase

Purpose: Database infrastructure, authentication, and data storage

Supabase provides our core data storage and user authentication services. All user data, including your account information, teaching materials, and conversation history, is stored securely through Supabase.

Data Processed: User account data, authentication credentials, teaching materials, documents, and conversation history

Data Location:

Primary storage: EU data centers
Read replicas: United States and Mumbai (India)

Cross-border transfer mechanism: EU Standard Contractual Clauses (SCCs), incorporated into Supabase's Data Processing Addendum.

References: Supabase Privacy Policy · Supabase DPA

Vercel

Purpose: Application hosting and performance analytics

Vercel hosts the TeamTeacher application and provides performance monitoring to ensure optimal service delivery. This includes application serving, edge caching, and basic analytics to monitor application health and performance.

Data Processed: Application usage patterns, performance metrics, and technical logs

Data Location: Global edge network; primary regions in the EU. Some technical logs may be processed in the United States.

Cross-border transfer mechanism: EU Standard Contractual Clauses (SCCs), covered under Vercel's Data Processing Addendum.

References: Vercel Privacy Policy · Vercel DPA

Analytics & Monitoring

PostHog

Purpose: Product analytics and user experience monitoring

PostHog helps us understand how educators use TeamTeacher, allowing us to improve features and user experience. We use this data to identify issues, optimize workflows, and develop new features that serve teachers better.

Data Processed: Anonymized usage patterns, feature interactions, and aggregated user behavior. We configure PostHog to minimize personal data collection.

Data Location: United States

Cross-border transfer mechanism: EU Standard Contractual Clauses (SCCs)

References: PostHog Privacy Policy · PostHog GDPR Compliance · PostHog DPA Generator

Langfuse

Purpose: LLM monitoring, governance, and quality assurance

Langfuse monitors our AI language model interactions to ensure quality, safety, and compliance. This includes tracking conversation quality, identifying potential issues, and maintaining responsible AI usage standards.

Important note: As part of this monitoring function, Langfuse processes AI prompt and response data — meaning the content you submit to TeamTeacher's AI features may be included in these logs. Langfuse does not use this data to train AI or ML models. If you have concerns about specific content, you can delete it at any time from your account settings.

Data Processed: AI conversation logs, prompt/response pairs, and model performance metrics

Data Retention: Logs are retained for up to one year, after which they are automatically deleted. Logs will be deleted sooner upon user request or account deletion.

Data Location: EU region data centers

No model training: Langfuse does not train or fine-tune ML/LLM models on client data. See Langfuse Privacy FAQ.

Cross-border transfer mechanism: Data remains within the EU/EEA. Langfuse DPA incorporates EU SCCs for any subprocessor transfers outside the EU.

References: Langfuse Privacy Policy · Langfuse DPA · Langfuse Security

Data Security Services

Skyflow

Purpose: PII de-identification and data privacy vault

Skyflow provides de-identification services that detect and redact personally identifiable information (such as names, email addresses, phone numbers, and ID numbers) from content before it is sent to AI providers or stored. This is a core part of our privacy-by-design approach for protecting sensitive information in teaching content.

Data Processed: User-submitted text content is scanned for PII. Detected PII tokens are replaced with anonymized placeholders. Original PII values are stored in Skyflow's secure vault for re-identification when needed.

Data Location: United States

Cross-border transfer mechanism: EU Standard Contractual Clauses (SCCs)

References: Skyflow Privacy Policy · Skyflow Security

Knowledge Base Services

Qdrant

Purpose: Vector database for knowledge base hosting and semantic search

Qdrant powers our knowledge base search capabilities, allowing AI agents to efficiently retrieve relevant information from curriculum documents and teaching resources. When you use agent tools that search knowledge bases, your search queries are processed by Qdrant.

Data Processed: Vector embeddings of knowledge base content, search queries, and retrieval results

Data Location: Google Cloud Platform, US East region (us-east1). Data is processed in the United States.

Cross-border transfer mechanism: EU Standard Contractual Clauses (SCCs), incorporated into Qdrant's Cloud Data Processing Agreement.

References: Qdrant Privacy Policy · Qdrant DPA · Qdrant Security

AI Language Model Providers

TeamTeacher integrates with multiple AI language model providers to offer you choice and flexibility in your teaching assistance. Depending on which agent and model you select, your conversation may be processed by one of these providers.

Commitments that apply to all AI providers:

No model training on your data. None of our AI providers use your prompts, conversations, or content to train or fine-tune their models. We maintain no-training agreements with all AI providers we engage.
Limited retention for operational purposes only. AI providers may retain submitted content for a short period (typically up to 30 days) solely for safety monitoring and abuse prevention, after which it is deleted, except as required by law. This content is not used for model training.
Processor-only relationship. AI providers act solely as data processors on our behalf. They may not use your data for any purpose other than returning a response to your request.

Anthropic

Anthropic provides advanced language models (Claude) with strong reasoning capabilities, extended context windows, and robust safety features. Used for complex lesson planning, detailed content analysis, and sophisticated teaching assistance.

Data Location: United States

Cross-border transfer mechanism: EU Standard Contractual Clauses (SCCs)

References: Anthropic Privacy Policy

OpenAI

OpenAI offers versatile language models with broad capabilities across various teaching tasks, as well as image generation and document embeddings. Used for text and image generation, conversation, and general teaching assistance.

Data Location: United States

Cross-border transfer mechanism: EU Standard Contractual Clauses (SCCs)

References: OpenAI Privacy Policy · OpenAI Data Processing Addendum

Google (Gemini)

Google provides language models with strong performance on educational tasks, including multimodal reasoning and image generation. Used for text and image generation, long-context document analysis, and cost-effective teaching assistance.

Data Location: United States

Cross-border transfer mechanism: EU Standard Contractual Clauses (SCCs) via Google Cloud DPA.

References: Google AI Privacy Policy · Google Cloud DPA

Mistral AI

Mistral provides efficient language models with strong multilingual capabilities. Used for text generation, coding assistance, and general teaching tasks.

Data Location: European Union

Cross-border transfer mechanism: Data processed within the EU/EEA.

References: Mistral Privacy Policy

OpenRouter

OpenRouter provides a unified API gateway to a range of additional language models not otherwise listed above. OpenRouter is used only when you explicitly select an OpenRouter-hosted model for your conversation — your content is never routed through OpenRouter by default. Content may be processed by AI providers beyond those listed individually on this page. See our models page for the full list of available models and their providers.

Data Location: United States

Cross-border transfer mechanism: EU Standard Contractual Clauses (SCCs)

References: OpenRouter Privacy Policy

Perplexity AI

Perplexity provides AI-powered web search with cited sources. Used for real-time research, fact-checking, and retrieving current information to support teaching tasks.

Data Location: United States

Cross-border transfer mechanism: EU Standard Contractual Clauses (SCCs)

References: Perplexity Privacy Policy

Tavily

Tavily provides AI-optimized web search. Used for the general web search tool, retrieving current information and web content to support teaching tasks.

Data Location: United States

Cross-border transfer mechanism: EU Standard Contractual Clauses (SCCs)

References: Tavily Privacy Policy

Data Processing Standards

All subprocessors are required to:

Process data only as instructed by TeamTeacher
Maintain appropriate technical and organizational security measures
Comply with applicable data protection regulations (GDPR, FERPA, etc.)
Support data subject rights (access, deletion, portability)
Notify TeamTeacher of any data breaches promptly
Delete or return data upon contract termination
Not use TeamTeacher data to train AI or ML models (where applicable)

Changes to Subprocessors

We will update this page when we add, replace, or remove subprocessors.

For existing features and services: We will provide at least 30 days' advance notice before adding or replacing a subprocessor in a way that affects currently available features. Notice will be given by updating this page and sending an email to the address registered on your account. Schools and districts with an active DPA may object to a proposed subprocessor change within this 30-day window by contacting privacy@teamteacher.ai.

For new or experimental features (alpha/beta): New features clearly marked as alpha, beta, or experimental may introduce new subprocessors as part of that feature's specific processing. Because these features require explicit opt-in, your decision to enable them constitutes acknowledgment of the associated processing. Subprocessors introduced exclusively for opt-in features will be listed here at or before the time of general availability.

Questions

If you have questions about our subprocessors or data processing practices, please contact us at privacy@teamteacher.ai.

Last updated: April 16, 2026