1. OUR COMMITMENTS (SUMMARY)
For every AI subprocessor that processes your Inputs or Outputs: (a) the provider is bound by contract to use your data only to provide the service to us; (b) the provider is prohibited from using your Inputs or Outputs to train its models, or we have enabled the provider's zero-retention or no-training configuration, as reflected in our agreements; (c) cross-border transfers are protected by the safeguards described in the Privacy Policy; and (d) we do not use your Inputs or Outputs to train generative AI models without your prior express opt-in consent, and we never use investor suitability data or KYC files for AI training.
Hutfin's primary conversational and analytical AI capability, Vera 2.0, is built and operated in-house by Hutfin rather than licensed from a third party. Where Vera 2.0 is used, your Inputs and Outputs are processed within Hutfin's own infrastructure rather than transmitted to an external model provider, and the commitments above are enforced directly by Hutfin. Sections 2 and 6.6 below describe where third-party providers are used for supporting functions.
2. CURRENT AI SUBPROCESSORS
The list below reflects Hutfin's currently engaged providers as of the Last Updated date above. Where a function is performed in-house rather than by a third party, that is stated explicitly. Categories without a named third-party provider below are either performed in-house, not yet in active use, or will be added here once a provider is engaged; this page is updated on a rolling basis as that happens.
Large Language Model Inference Provider: Vera 2.0, Hutfin's in-house-built large language model. Vera 2.0 is developed and operated by Hutfin and is not a third-party subprocessor. Function: generates responses for AI chat, document question-answering, drafting, and analysis features. Processing location: Hutfin's own hosting infrastructure, as described in the Privacy Policy. Retention: because Vera 2.0 is operated in-house rather than by an external provider, your Inputs and Outputs are not transmitted to a third-party model vendor for inference. The no-training-without-opt-in commitment in Section 6.4 of the Privacy Policy applies in full, and is enforced directly by Hutfin rather than through a third-party contract.
Cloud Hosting and Infrastructure Function: hosts Platform services, AI orchestration, and storage. Retention: per Hutfin's retention schedule in the Privacy Policy.
Vector Database and Embeddings Function: stores document embeddings that power AI search and question-answering. Retention: embeddings deleted when the underlying content is deleted.
Safety and Content Classification Function: screens Inputs and Outputs for fraud, abuse, prompt-injection, and prohibited content. Retention: safety-flagged records retained per Section 6.12 of the Privacy Policy.
Transcription (where meeting-notes or call features are offered) Function: converts audio to text for features you invoke, with notice. Retention: transient processing; transcripts stored by Hutfin per the Privacy Policy.
Identity and Liveness AI (KYC) Provider: Plaid Inc. (Plaid Identity Verification). Function: document authenticity, facial comparison, and liveness checks in KYC flows. Processing location: as described in Plaid's Privacy Policy, available at https://plaid.com/legal/#privacy-policy. Retention: per the biometric retention and destruction schedule in the Privacy Policy and applicable law, and per Plaid's own data handling terms.
Analytics and Monitoring Provider: Google (Google Analytics; reCAPTCHA for security) Function: usage analytics and security signals, as described in the Privacy Policy and Cookie Policy. These services do not receive the content of your Inputs or Outputs. Processing location: United States and global infrastructure. Retention: per provider settings described in the Privacy Policy (for example, up to 26 months for Google Analytics).
3. CHANGES TO THIS LIST
We may add, replace, or remove AI subprocessors. This page is the authoritative current list. Where a specific AI Feature routes data to a provider under materially different terms than those summarized above, we say so in that feature's interface before you use it. Material changes affecting how your personal information is processed are notified per the Privacy Policy.
4. QUESTIONS
Questions about AI subprocessors or AI privacy: privacy@hutfin.com (subject "AI Privacy"). Written correspondence: Hutfin Global Technologies, Inc., 1720 Stratton Dr, Virginia Beach, VA 23456, USA.