Annex A — Anthropic Platform Processing

Feature key: anthropic_platform
Legal feature name: Senex-managed Anthropic Platform Processing
DPA version: anthropic-platform-dpa-v2026-07-07
Viewable route: /legal/dpa/amendments/anthropic-platform-v2026-07-07
Provider: Anthropic
Provider role: Senex-appointed conditional Sub-processor for this feature
Not BYOK: This annex does not apply to Anthropic bring-your-own-key processing using Customer's own Anthropic account and API key.

1. Scope

This annex applies only where Customer affirmatively enables anthropic_platform and Senex uses a Senex-held Anthropic account, API key, workspace, or platform configuration to process Customer Data for Customer's authorised use of the Service.

This feature is not enabled by default and is not part of the Cloudflare-only default configuration.

2. Processing purpose

Customer instructs Senex to use Anthropic to provide artificial-intelligence model processing for authorised Service workflows, including where applicable:

1. answer generation;
2. question answering over Customer's corpus;
3. summarisation;
4. classification;
5. reasoning and drafting support;
6. extraction of structured information;
7. transformation of retrieved context into outputs;
8. citation-aware response generation; and
9. related debugging, safety, abuse-prevention, security, and operational processing necessary to provide the feature.

This annex does not authorise public web search, public web fetch, external source retrieval, or persistent ingestion of external public documents into Customer's corpus unless Customer has separately accepted the applicable Public Source Retrieval / public_web_search DPA amendment.

3. Customer Data that may be processed

For this feature, Senex may transmit to Anthropic the following categories of Customer Data, limited to what is reasonably necessary for the authorised workflow:

1. prompts, questions, system instructions, user instructions, and task metadata;
2. retrieved passages, document excerpts, chunks, tables, titles, filenames, and source metadata from Customer's corpus;
3. tenant, workspace, user, run, trace, request, and configuration metadata;
4. outputs, intermediate model responses, tool-call metadata, and citation metadata;
5. error, safety, abuse-prevention, rate-limit, latency, and operational metadata; and
6. other Customer Data intentionally included by Customer's authorised users in prompts or selected context.

Customer should not use this feature for protected health information, special-category personal data, criminal-offence data, children's data, legally privileged material, regulated client data, or similarly sensitive information unless the applicable Order Form, DPA, Anthropic contractual posture, and Customer's own compliance assessment expressly permit that use.

4. Anthropic legal and transfer posture

Senex will use this feature only through an Anthropic commercial/API arrangement that incorporates Anthropic's data-processing terms or equivalent written terms between Senex and Anthropic.

Customer authorises Senex to appoint Anthropic as a conditional Sub-processor for this feature. Anthropic may process Customer Data in the United States or other Anthropic service locations. To the extent this involves a Restricted Transfer of Customer Personal Data, the transfer is governed by the DPA, Anthropic's applicable data-processing terms, Standard Contractual Clauses, United Kingdom transfer terms, Swiss transfer terms, adequacy mechanisms, and supplementary measures where applicable.

Senex will not represent this feature as European-Union-only, Swiss-only, or Cloudflare-only processing.

5. Model training and provider use limitation

Senex will use an Anthropic commercial/API configuration under which Anthropic is not permitted to train models on Customer Content submitted through the commercial services unless Customer separately authorises otherwise in writing.

If Senex becomes aware of a material change to Anthropic's applicable terms that would permit model training on Customer Content without Customer's express permission, Senex must either disable the feature for affected Customers until re-authorised or obtain updated Customer authorisation.

6. Security and operational controls

Senex will implement reasonable controls designed to:

1. keep anthropic_platform disabled unless a valid Feature Consent and DPA Reference exist;
2. enforce the current DPA version and matching disclosure hash server-side;
3. minimise Customer Data transmitted to Anthropic to what is reasonably necessary for the authorised feature;
4. prevent transmission through this feature for tenants or workspaces classified as Cloudflare-only, protected-health-information, or otherwise restricted unless separately approved;
5. log non-content audit metadata for feature enablement and use;
6. protect Senex-held Anthropic credentials; and
7. disable future egress after revocation, stale terms, or operator disablement.

7. Customer responsibilities

Customer is responsible for:

1. deciding whether Customer Data is suitable for Anthropic processing;
2. ensuring that Customer has the lawful basis, notices, permissions, and regulatory authority to instruct this processing;
3. assessing whether additional confidentiality, sector-specific, or international-transfer safeguards are required;
4. validating outputs before reliance, publication, disclosure, or regulated use; and
5. managing authorised users and internal policies governing which content may be submitted to the feature.

8. Retention and deletion

Anthropic's retention of Customer Data for this feature is governed by Senex's agreement with Anthropic and Anthropic's applicable commercial/API data-processing terms. Unless an Order Form expressly states otherwise, Senex does not promise Zero Data Retention for this feature.

Disabling or revoking this feature stops future egress to Anthropic but does not automatically delete outputs, audit records, logs, backups, or Customer corpus records already created in Senex, or records retained by Anthropic as permitted under the applicable Anthropic terms, law, security, abuse-prevention, or audit requirements.

9. Activation acknowledgement text

Customer-facing activation must include substantially the following acknowledgement:

Senex may send Customer Data to Anthropic using Senex's Anthropic API account for artificial-intelligence model processing, including prompts, retrieved context, document excerpts, request metadata, and outputs. Anthropic is a Senex-appointed conditional Sub-processor for this feature and may process data in the United States or other Anthropic service locations under Anthropic's commercial data-processing terms and applicable transfer safeguards. This is not bring-your-own-key processing, is not Cloudflare-only processing, and is not enabled by default.

Customer must affirmatively accept the current version of this annex before enablement.