# Addendum Claude

> Source: https://bryter.com/addendum-claude/

[#](#)

[Download PDF](https://bryter.com/wp-content/uploads/2026/05/260522_Addendum_Claude_DRAFT_ENG.pdf)

# Addendum Claude

**ADDENDUM TO THE DATA PROCESSING ADDENDUM**

**Use of Anthropic Claude Models in BEAMON AI**

between

Customer as defined in the applicable Order (hereinafter “**Customer**”)

and

BRYTER GmbH, Biebergasse 2, 60313 Frankfurt am Main, Germany (hereinafter “**BRYTER**”)

(together also referred to as the “**Parties**” and each also referred to as a “**Party**”).

This addendum (“**Addendum**”) supplements the Data Processing Addendum entered into between the Parties in its respective version (“**DPA**”) and sets out the additional data protection terms applicable where Customer uses, as part of BEAMON AI, large language models provided by Anthropic PBC.

## 1. Definitions

1.1        “**Anthropic**” means Anthropic PBC, 548 Market Street, PMB 90375, San Francisco, CA 94104, USA.

1.2        “**Claude Models**” means the large language models developed and operated by Anthropic that Customer may use through BEAMON AI.

1.3        “**DPF**” means the adequacy decision of the European Commission of 10 July 2023 on the EU-U.S. Data Privacy Framework (Decision (EU) 2023/1795).

1.4        All other capitalised terms shall have the meaning set out in the DPA, the MSA and the Definitions (Appendix 1 of the MSA).

## 2. Scope

2.1        This Addendum applies only for as long as Customer has activated the use of Claude Models within BEAMON AI. Activation is performed by Customer via the designated function of the Software (opt-in).

2.2**           **This Addendum supplements the DPA and the MSA. In the event of any conflict, the provisions of this Addendum shall prevail over the DPA; otherwise, the DPA and the MSA remain in full force and effect.

## 3. Extension of the List of Sub-Processors

3.1        Upon activation pursuant to Section 2.1, Anthropic PBC becomes an additional Sub-Processor within the meaning of Section 9 DPA. The list of Sub-Processors set out in Schedule 1 of the DPA is supplemented by the entry contained in Annex A to this Addendum.

3.2       Processing by Anthropic is limited to providing the inference functionality of the Claude Models (processing of prompts and generation of outputs). No further processing, in particular for the purposes of improving or training the models, takes place (Section 12.2 MSA).

3.3**           **The Claude Models are made available technically through Microsoft Foundry. Microsoft solely provides the API infrastructure and does not process the content of prompts and outputs.

## 4. Processing Outside the European Union

4.1        Notwithstanding Section 10 DPA, the processing of Personal Data by Anthropic in connection with the provision of the Claude Models takes place wholly or partly outside the European Union and the European Economic Area, in particular in the United States of America.

4.2       The transfer is based on the DPF (Art. 45 GDPR). Anthropic is certified under the DPF. The certification can be verified at https://www.dataprivacyframework.gov.

4.3       Should the DPF become wholly or partly invalid or should Anthropic's DPF certification end, the EU Standard Contractual Clauses (Module 3: Processor-to-Processor, Implementing Decision (EU) 2021/914) shall apply between BRYTER and Anthropic as appropriate safeguards within the meaning of Art. 46(2)(c) GDPR.

## 5. Professional Secrets and Professional Confidentiality

5.1        Anthropic has not entered into specific contractual undertakings vis-à-vis BRYTER in respect of professional confidentiality obligations within the meaning of Sections 43a, 43e BRAO, Section 18 BNotO or Section 203 of the German Criminal Code (StGB). Compliance with professional confidentiality obligations when using BEAMON AI with Claude Models is therefore not assured insofar as Customer processes Professional Secrets within the meaning of Section 8 DPA.

5.2       It is Customer's responsibility to assess, prior to processing any Professional Secrets via Claude Models, whether the intended use is compatible with Customer's professional obligations. For the processing of Professional Secrets, Customer may continue to use the other functions of the BRYTER Software within the framework of the DPA without restriction.

5.3       Customer shall indemnify BRYTER against all third-party claims, damages, fines and costs incurred by BRYTER as a result of the processing of Professional Secrets through Claude Models contrary to the notices in this Section 5.

## 6. Obligations of BRYTER

6.1        BRYTER has entered into a data processing agreement with Anthropic that meets the requirements of Art. 28(4) GDPR. Such agreement will be made available to Customer upon request.

6.2       Where Anthropic fails to comply with its data protection obligations, BRYTER remains liable to Customer in accordance with Section 9.4 DPA.

6.3       BRYTER will inform Customer without undue delay if (i) Anthropic's DPF certification ends, (ii) the DPF is declared wholly or partly invalid, or (iii) any other circumstances material to the lawfulness of the transfer occur.

## 7. Deactivation and Termination

7.1        Customer may revoke the activation pursuant to Section 2.1 at any time via the designated function of the Software. Upon revocation, this Addendum ceases to apply. Processing that has already taken place remains unaffected.

7.2       In the cases set out in Section 6.3, Customer is entitled to discontinue the use of the Claude Models with immediate effect. Any other termination rights of Customer in respect of the MSA and the DPA remain unaffected.

7.3       Termination of this Addendum does not affect the continued existence of the MSA, the applicable Order or the DPA. Customer may continue to use BEAMON AI without the Claude Models, to the extent other models are available in BEAMON AI.

## 8. Final Provisions

8.1        To the extent Section 9.6 DPA (special rule for AWS and Microsoft Azure) applies, it shall apply accordingly to Anthropic, with the proviso that an objection by Customer against Anthropic as a Sub-Processor shall not entitle Customer to terminate the MSA but shall only lead to deactivation pursuant to Section 7.1.

8.2       In all other respects, the final provisions of the DPA and the MSA shall apply accordingly.

**Annex A**

-

**Supplement to Schedule 1 DPA (Sub-Processors)**

The list of Sub-Processors set out in Schedule 1 of the DPA is supplemented, upon activation of the Claude Models pursuant to Section 2.1 of this Addendum, by the following entry:

| **Sub-Processor** | **Service Provided** | **Corporate Location** | **Server Location** |
| --- | --- | --- | --- |
| Anthropic PBC | Provision of Claude Models (LLM inference for BEAMON AI) | 548 Market Street, PMB 90375, San Francisco, CA 94104, USA | USA (Anthropic-managed infrastructure on AWS and/or GCP) |

The transfer is based on the EU-U.S. Data Privacy Framework (Art. 45 GDPR); the EU Standard Contractual Clauses apply subsidiarily pursuant to Section 4.3 of this Addendum.

In connection with the use of the Claude Models, the entry "Microsoft Azure" in Schedule 1 of the DPA covers solely the API infrastructure provided by Microsoft through Foundry (routing, authentication, endpoint management). The substantive processing of prompts and outputs is not carried out by Microsoft but exclusively by Anthropic in accordance with the entry above.