Data Processing Agreement.

This Data Processing Agreement (“DPA”) forms part of the Terms of Service (or other written agreement) between Zentrr, Inc. (“Zentrr”) and the customer entity that accepted them (“Customer”), and applies whenever Zentrr processes Personal Data contained in Customer Content on Customer's behalf.

For that processing, Customer is the controller (or a processor acting on behalf of its own customers) and Zentrr is the processor (or sub-processor). Each party will comply with its own obligations under applicable Data Protection Laws. In the event of a conflict between this DPA and the Terms with respect to the processing of Personal Data, this DPA prevails. Enterprise customers may request a countersigned copy of this DPA (with the Standard Contractual Clauses appended) by contacting privacy@zentrr.com.

“Data Protection Laws” means all laws applicable to the processing of Personal Data under the agreement, including (as applicable) the EU General Data Protection Regulation 2016/679 (“GDPR”), the UK GDPR and Data Protection Act 2018, the Swiss Federal Act on Data Protection (“FADP”), and US state privacy laws including the California Consumer Privacy Act as amended by the CPRA (“CCPA”).

“Personal Data” means information relating to an identified or identifiable natural person that is contained in Customer Content and processed by Zentrr on Customer's behalf.

“Security Incident” means a breach of security leading to the accidental or unlawful destruction, loss, alteration, unauthorized disclosure of, or access to Personal Data processed by Zentrr.

“Sub-processor” means a third party engaged by Zentrr to process Personal Data on Customer's behalf.

“Customer Content,” “Output,” and “Agent” have the meanings given in the Terms. “Controller,” “processor,” “data subject,” and “processing” have the meanings given in the GDPR.

Zentrr will process Personal Data only on Customer's documented instructions, including: (a) to provide, secure, and support the Service as described in the Terms; (b) as configured by Customer and its users in the product (knowledge, integrations, agent access); and (c) as further instructed in writing (including support requests). Zentrr will inform Customer if, in its opinion, an instruction infringes Data Protection Laws (and may suspend the affected processing until resolved).

The subject matter, duration, nature and purpose of processing, the types of Personal Data, and the categories of data subjects are set out in Annex A below.

No AI training. Zentrr does not use Customer Content (including Personal Data) to train any AI model. Inference runs on AWS Bedrock; per AWS, prompts and completions are not used to train or improve foundation models. By default, model providers have no access to them.

Optional models that share data with Anthropic. Certain Anthropic models served via AWS Bedrock are offered only under provider data sharing and are disabled by default. Where an authorized user of Customer explicitly enables such a model on a specific agent, Customer instructs Zentrr (as a documented processing instruction, recorded in Customer's audit log) to permit sharing of that agent's prompts and completions with the model provider (Anthropic), retained by the provider for up to thirty (30) days for trust & safety and abuse detection only — never for model training. This applies solely to the opted-in agent. Customer remains responsible for assessing the suitability of such a model for any Special Categories of Personal Data or other regulated data it processes.

Taking into account the state of the art and the nature of the Personal Data processed, Zentrr implements and maintains the technical and organizational measures set out in Annex B (the “Security Measures”), including encryption at rest (AES-256, with keys held in a managed key-management service) and in transit (TLS 1.2 or higher), per-tenant isolation enforced at the row level by the database itself alongside application-level scoping, and audit logging. Zentrr may update the Security Measures from time to time, provided the updates do not materially reduce the overall protection of Personal Data.

An overview of our security and control measures is published at zentrr.com/#controls.

Customer provides general authorization for Zentrr to engage Sub-processors. The current list — including purpose, location, and category for each — is published at zentrr.com/legal/subprocessors and is incorporated into this DPA.

Zentrr will provide at least thirty (30) days' notice before adding or replacing a Sub-processor that materially handles Customer Content (via email to account admins or in-product notice). Customer may object on reasonable, documented data-protection grounds within that window; if the parties cannot resolve the objection, Customer may terminate the affected portion of the Service and receive a pro-rata refund of prepaid fees for the unused period.

Zentrr imposes data-protection obligations on each Sub-processor that are no less protective than this DPA and remains responsible for each Sub-processor's performance.

Third-party channels and integrations that Customer chooses to connect (QuickBooks Online, Google Drive, a website, an S3 bucket, Slack, etc.) act on Customer's instructions under Customer's own agreements with those providers; they are not Zentrr Sub-processors.

Zentrr will notify Customer without undue delay, and in any event within seventy-two (72) hours, after becoming aware of a Security Incident affecting Customer's Personal Data. The notification will describe, to the extent known: the nature of the incident, the categories and approximate volume of Personal Data and data subjects concerned, the measures taken or proposed to address it, and a contact point for further information.

Zentrr will take reasonable steps to contain and remediate the incident and will provide information reasonably required for Customer to meet its own notification obligations. Zentrr's notification of a Security Incident is not an acknowledgment of fault or liability.

Zentrr stores Customer Content in the United States. Primary storage and application processing run in AWS us-east-1; encrypted backup copies are kept in AWS us-west-2, in a separate Zentrr-controlled AWS account, for disaster recovery. An Enterprise deployment that is BYOC, in Customer’s own AWS account, keeps Customer Content in that account and in the AWS region it runs in, and the us-west-2 backup copies do not apply to it. Model inference runs on AWS Bedrock in US regions while the per-workspace “keep model calls in the United States” setting is on, which it is by default and must be while a Business Associate Agreement under § 13 is in force; if Customer turns it off, inference may be routed to other AWS regions. Where Data Protection Laws of the EEA, the United Kingdom, or Switzerland apply to a transfer of Personal Data to Zentrr in the United States:

(a) the EU Standard Contractual Clauses approved by Commission Implementing Decision (EU) 2021/914 (“SCCs”), Module Two (controller → processor) or Module Three (processor → processor) as applicable, are incorporated into this DPA by reference, with Customer as data exporter and Zentrr as data importer; Annexes I and II of the SCCs are populated by Annex A and Annex B of this DPA; the optional docking clause applies; the governing law and forum are those of Ireland;

(b) for UK transfers, the SCCs apply as amended by the UK International Data Transfer Addendum issued by the ICO; and (c) for Swiss transfers, the SCCs apply with the adaptations required by the FDPIC, with “Member State” read to include Switzerland and data subjects able to enforce their rights in Switzerland.

Where the CCPA or another US state privacy law applies, Zentrr acts as Customer's “service provider” or “processor.” Zentrr will not: (a) sell or share Personal Data; (b) retain, use, or disclose Personal Data for any purpose other than performing the Service under the agreement or as permitted by law; (c) retain, use, or disclose Personal Data outside the direct business relationship with Customer; or (d) combine Personal Data with personal information from other sources except as permitted for service providers.

Zentrr certifies that it understands these restrictions and will comply with them, will notify Customer if it can no longer do so, and grants Customer the right to take reasonable steps to stop and remediate unauthorized use of Personal Data. See also Do Not Sell or Share.

Zentrr will make available information reasonably necessary to demonstrate compliance with this DPA, including its security documentation, sub-processor list, compliance posture reports, and — once issued — its SOC 2 Type II report. The parties agree these materials satisfy Customer's audit rights in the first instance.

Where Data Protection Laws grant a non-waivable audit right, Customer (or its mandated independent auditor, not a Zentrr competitor) may audit Zentrr's compliance with this DPA no more than once per twelve (12) months, on at least thirty (30) days' written notice, during business hours, subject to reasonable confidentiality and security requirements, at Customer's expense, and without access to other customers' data or to Sub-processor facilities (Sub-processor compliance is demonstrated through their own certifications and reports).

During the term, including the thirty (30) days before the closing date of a workspace Customer schedules for closing, a Workspace Admin can download a copy of Customer Content through the Service's built-in workspace export, from Settings at any time; the download names anything it does not yet hold and where in the Service to take it from. Customer can request a copy of any other Customer Content by contacting privacy@zentrr.com. That is how Personal Data is returned to Customer: there is no export period after the closing date, and none for a workspace Customer chooses to close immediately. Upon termination or expiration, Zentrr deletes Personal Data on these schedules: account data stops being reachable through the Service when a deletion request is honored; chat messages are retained for the period set by Customer's plan, as are activity audit records; administrative and compliance audit records are retained for at least one (1) year and are not on an automatic deletion schedule, all as described in the Privacy Policy.

Residual copies persist for a limited time in the storage layer beneath the Service — document storage retains a prior version of a deleted file for up to three hundred sixty-five (365) days, and encrypted backups are retained on a rolling schedule — after which they are destroyed automatically. Residual copies are not reachable through the Service. Zentrr may retain Personal Data to the extent required by law, subject to the protections of this DPA.

Subject matter & duration: processing of Personal Data contained in Customer Content to provide the Zentrr Service, for the term described in § 15.

Nature & purpose: hosting and storage; AI inference to generate Outputs (drafts, summaries, packets, replies) for human review; Knowledge syncing and citation; routing sign-off requests to Approvals; channel delivery as configured by Customer; audit logging; support.

Categories of data subjects: Customer's users and personnel; Customer's clients, patients, policyholders, or other end customers; and other individuals whose Personal Data appears in Customer Content.

Types of Personal Data: identification and contact data; professional and employment data; business, financial, insurance, or other records Customer connects or uploads; message content; usage and approval metadata. Special categories only if and to the extent Customer chooses to submit them (PHI requires a BAA per § 13).

Frequency: continuous, as driven by Customer's use. Retention: per § 12.

Encryption: TLS 1.2 or higher in transit; AES-256 at rest for the database and customer document storage, with keys held in a managed key-management service.

Tenant isolation: per-organization data partitioning enforced at the row level by the database itself, alongside application-level scoping; per-organization Knowledge stores. On Enterprise there are two ways to deploy: a single-tenant deployment operated by Zentrr, with Customer’s data in Zentrr’s AWS; or BYOC, in Customer’s own AWS account.

Access control: role-based access (Workspace Admin, Team Lead, Member) backed by a managed identity service with MFA support; least-privilege access controls on the cloud infrastructure; production access restricted to authorized personnel.

Network & infrastructure: private networking, with no direct access from the public internet, for the database and the web application’s servers; managed AWS services (HIPAA-eligible); a web application firewall at the edge; storage and application processing in the United States, with model-call region pinning available as a per-workspace setting (see § 09).

Logging & monitoring: audit logging of administrative API calls on the cloud infrastructure; application-level tamper-evident audit trail of chats, citations, and approvals, in which a chat turn is recorded as a digest of its content rather than its text; email an Agent sends with its email and calendar tools, and email an app sends, is recorded with its recipient count, the recipients’ domains and a digest of its subject and body, not the addresses or the text; what Zentrr support staff open through a support window Customer grants is recorded, naming the window; alerting on anomalous activity.

AI-specific measures: inference confined to AWS Bedrock, with no Customer Content sent to a third-party AI provider by default — the one exception is a small set of models AWS Bedrock serves only under provider data sharing, which are off unless a workspace administrator opts a specific Agent into one, an action recorded in the workspace audit log; no training of any model on Customer Content; platform content-safety screening on every turn — violence, hate, sexual content, insults, and misconduct filters, in both directions — plus prompt-injection screening of input from anyone who is not an authenticated member of the Customer’s workspace, with social security and payment card numbers blocked in both directions; human sign-off required in Approvals before Agent actions with external effect.

Resilience & lifecycle: multi-AZ database, encrypted backups on a standard rotation cycle, documented incident response, vendor (Sub-processor) review, and personnel confidentiality and training per § 04.