Data processing agreement

Last updated: 11 October 2026 · Version 2.0

This data processing agreement ("DPA") is part of the agreement between you, the customer, and Smidt Beheer, trading as Recall ("Recall", "we", "us"), Binnendelta 11M, 1261 WZ Blaricum, the Netherlands (KvK 99124254), under our terms of service. It sets out how we process personal data on your behalf, as Article 28 of the General Data Protection Regulation (GDPR) requires.

1. Scope

2. Roles

You are the controller of Customer Personal Data and we are your processor. If you process the data as a processor for someone else, you confirm that your controller has authorised you to engage us, and we are your sub-processor. Annex 1 describes the processing.

3. Instructions

4. Confidentiality

Only people who need access to run, support or secure the service can access Customer Personal Data, and each of them is bound by a duty of confidentiality, by contract or by law. They access it only to give support you ask for, to investigate abuse or a security incident, or to comply with the law.

5. Security

6. Sub-processors

7. International transfers

8. Requests from authorities

If an authority asks us for Customer Personal Data, we refer it to you where we can. We disclose data only when a law that binds us obliges us to, after checking the request, and we challenge requests that we consider unlawful. We disclose only what is required and tell you about the request, unless the law forbids that.

9. Assistance

10. Personal data breaches

11. Deletion and return

12. Information and audits

13. Your obligations

14. Liability, precedence and law

Annex 1: Processing details

Subject matterProviding Recall, memory for AI agents, to you.
Nature of the processingStoring and organising the messages and files you send; analysing them with AI models to extract facts, write summaries and peer cards and compute embeddings; searching and answering questions about them; merging and updating memory in the background; sending data to your systems (API responses, webhooks you set up, MCP clients you connect); exporting and deleting data.
PurposeTo provide the service under the terms of service.
DurationFor the term of the agreement, and until the data is deleted under section 11.
Data subjectsThe people whose conversations or data you send, typically your end users, your staff and people mentioned in conversations.
Types of personal dataWhatever you send: message and file content, the names and identifiers you choose for peers, sessions and workspaces, metadata, and the facts, peer cards, summaries and embeddings derived from them. Special categories of data only if you send them (section 13).
FrequencyContinuous, as you use the service.
RetentionUntil you delete the data or your account, as described in the privacy policy.
LocationStorage and application servers in Frankfurt, Germany (EU). Processing by AI model providers in the United States (Annex 3).

Annex 2: Security measures

Access to the service

Separation

Encryption

Data minimisation

Organisation and operations

Annex 3: Sub-processors

Sub-processorWhat it doesLocation
Vercel Inc.Runs the API, dashboard and background work, and serves the websiteFrankfurt, Germany (EU); company in the USA
NeonDatabase: stores all Customer Personal DataFrankfurt, Germany (EU); company in the USA
OpenRouter, Inc.Passes requests to AI models for text generation and embeddingsUSA
OpenAI (through OpenRouter)Runs the AI models that build memory and answer questions, and computes embeddingsUSA
Anthropic, PBC (through OpenRouter)Runs the AI models that build memory and answer questionsUSA

These providers do not receive Customer Personal Data and are therefore not sub-processors; we list them for transparency: Resend (sign-in and account emails to account members), Stripe Payments Europe, Ltd., Ireland (payments, invoices and tax), Cloudflare (DNS) and GitHub (an automated uptime check of our health endpoint).