Data Processing Agreement (DPA)
Version 5 of November 16, 2026
Between
The Controller ("the Institution"): the educational institution (or its governing body, or the competent education authority) on whose behalf the teacher using Examino acts in the course of their duties.
Name:
Address:
Represented by:
DPO contact:
And:
The Processor ("Examino"): EXAMINO, a French simplified joint-stock company (SAS) with a share capital of EUR 1,000, whose registered office is at 49 RUE DE PONTHIEU 75008 PARIS, France, registered with the Paris Trade and Companies Register under number 989 510 375, represented by Mr. Dimitri Nicolas, acting with full authority as President. DPO contact: dpo@examino.ai
Preamble
Examino publishes an application that assists with grading student work and supports teaching. The teacher, an employee or agent of the Institution, uses it in the course of their duties to grade student work and, depending on the features enabled, to manage classes, share resources and assessments with students, and return their graded work to them. In line with the doctrine of the CNIL (French data protection authority), the Institution is the controller and Examino the processor within the meaning of Article 28 GDPR for student data. Examino acts as controller for teachers' account, billing and commercial relationship data, which are governed by its privacy policy.
This agreement is entered into unilaterally by Examino for the benefit of the Institution as of the first processing of the Institution's student data (import of student work, creation of a class or student list, first login of a student). The Institution may ratify it by countersigning; failing that, continued use after notice constitutes acceptance of these processing terms. Opening student accounts at institution level, in particular through the GAR (French resource access manager) or single sign-on (SSO), requires the Institution to sign this agreement. It prevails over any conflicting provision of the Terms of Use as regards the processing of student data.
Where the service is accessed through the Gestionnaire d'accès aux ressources (GAR), the commitments of this agreement are in addition to those of the GAR contractual framework, which prevails in the event of conflict.
1. Subject matter, nature and purpose of the processing
Subject matter | Assistance with grading exam papers and, where the teacher or the Institution enables them, related teaching tools: class and student management, student area, sharing of resources and online assessments, return of graded work to students. |
Nature of the operations | For grading: collection (import or scanning), linking the paper to the student, automatic redaction of the student's name, transcription (OCR), analysis by generative AI models, delivery of a proposed grading, review by the teacher (including by voice dictation), statistical harmonization of grades, suggestion of remediation exercises, export. For class management: creation and import of student lists (manual entry, file, synchronization from the Institution's digital workspace (ENT), the GAR or learning management system (LMS)), transmission of validated grades to the LMS at the teacher's request. For the student area: creation and authentication of student accounts, provision of the resources, assessments and graded work published by the teacher, collection of students' answers and work, collaborative document editing. For all of the above: storage, indexing for search, deletion. |
Purpose | To propose to the teacher a grading and comments, which the teacher reviews and validates under their sole pedagogical responsibility; to enable the teacher to organize classes, share teaching content and assessments with students, and return their results to them. In the current state of the service, students have no conversational interface with an AI model: they only access content published by their teacher, including AI-generated content that the teacher has validated or edited. Any change in this respect will be notified to the Institution in advance. |
Duration | Term of the teacher's subscription, within the retention periods of Article 8. |
2. Categories of data and data subjects
Data subjects: pupils and students of the Institution (including minors), whether or not they have a student area account; the teacher user and, where applicable, other Institution staff using the service.
Data processed:
student identification: last name, first name, class or group and, depending on the source, email address and identifier assigned by the Institution, the ENT, the GAR or the LMS;
student area connection data: session and device identifiers, timestamps, IP address (security logs);
schoolwork: content of papers, answers to online assessments, documents created or edited by the student;
results: grades, scores, comments and feedback proposed then validated, harmonized grades;
the teacher's teaching content: exam subjects, grading schemes, resources;
for the teacher: account data, edit history and, when using voice dictation, the recording of their voice sent for transcription.
Minimization: when papers are imported, Examino automatically masks the detected student last name and first name on the image. The redacted version is used for most processing by AI models; the original version is only sent to the step that detects the elements to mask and the page orientation. Other identifying marks that may be present (signature, handwritten notes) are not masked automatically, and papers imported as spreadsheets are not redacted. The student's identity taken from class lists is never sent to AI models. Wherever possible, Examino grades the digital transcription rather than the handwritten image (neutralizing handwriting).
These measures amount to pseudonymization of the data sent to AI models, not anonymization: Examino keeps the link between each paper and the student so that the teacher can use the results (details and limitations: redaction technical sheet, available on request).
Resources uploaded by the teacher to their library are not redacted: the teacher undertakes not to upload any data identifying a student there. A document created by a student is never used as the source of AI processing.
Sensitive data (Art. 9): no collection is intended; the free content of a paper, answer or document may incidentally reveal such data, and Examino makes no use of it. The teacher does not upload to the service any document relating to students' health or disability (individual support or care plans such as PAP, PAI, PPS), nor any mention of exam accommodations.
3. Documented instructions
Examino processes the data only on documented instructions from the Institution, as set out in this agreement and in the teacher's functional use of the application. Examino informs the Institution if, in its opinion, an instruction infringes the GDPR.
Examino does not use student data or imported content to train AI models, nor for any purpose of its own other than providing, securing and billing the service. The quality of the AI models used by Examino is evaluated exclusively on fictitious content or content whose use for this purpose has been expressly authorized.
4. Confidentiality and personnel
Persons authorized to process the data are bound by a contractual confidentiality obligation. Access is limited to authorized personnel on a least-privilege basis (RBAC).
Examino personnel only access an account's data for support purposes, at the user's request, or for the security of the service; staff access is logged.
5. Security (Art. 32)
Main measures:
encryption in transit (TLS 1.2+); encryption at rest at the hosting providers' level (AES-256, keys managed by the providers);
persistent storage hosted in the European Union;
role-based access control (least privilege); Examino personnel with access to critical systems or user data authenticate with multi-factor authentication (physical security key or authenticator app) and receive cybersecurity training;
logging of changes made to papers and gradings (timestamped history); authentication events of Examino's centralized account management system are logged and retained, and administrative changes to production systems are traceable through the providers' audit logs;
logical separation of customers: each institution's data is segregated at application level (tenant identifiers + role-based access control) on shared infrastructure; no dedicated infrastructure per institution is provided; students only access content published by their teacher and their own work and results;
student authentication delegated to the Institution's ENT, GAR or LMS where used; otherwise, access by class code and registered device, with no email address required;
managed redundant backups.
6. Subprocessors
The Institution gives general authorization for the use of the subprocessors listed below. Examino informs the Institution of any change (objection period: 30 days) and imposes on each subprocessor obligations equivalent to those set out herein. The up-to-date, detailed version of this list (precise location, transfer mechanisms) is published at: https://examino.ai/en/legal/subprocessors.
The Institution's ENT, GAR and LMS are not Examino subprocessors: data exchanges with these services (student lists, grades) take place at the initiative of the Institution or the teacher.
Infrastructure subprocessors:
Provider | Role | Processing or storage location | Safeguards |
|---|---|---|---|
Backblaze, Inc. (United States) | File storage | Netherlands (EU) | DPA, SOC 2 Type II |
Sendinblue, SAS (France) | Email delivery | Belgium (EU) | DPA, ISO 27001 |
Cloudflare, Inc. (United States) | Network, security and content delivery | Worldwide | DPA, DPF, ISO 27001, SOC 2 Type II |
Hetzner Online, GmbH (Germany) | Server hosting | Germany (EU) | DPA, ISO 27001 |
Intercom R&D, Unlimited Co. (Ireland) | Customer support | Ireland (EU) | DPA, ISO 27001, SOC 2 Type II |
OVH, SAS (France) | Server hosting | France (EU) | DPA, ISO 27001, SOC 2 Type II |
Pipedrive, OÜ (Estonia) | Customer relationship management | Ireland (EU), Germany (EU) | DPA, ISO 27001, SOC 2 Type II |
PlanetScale, Inc. (United States) | Database hosting | Germany (EU) | DPA, ISO 27001, SOC 2 Type II |
PostHog, Inc. (United States) | Analytics and technical monitoring | Germany (EU) | DPA, SOC 2 Type II |
Stripe Payments Europe, Ltd (Ireland) | Payments and billing | Worldwide | DPA, DPF, ISO 27001, SOC 2 Type II |
Trigger.dev, Inc. (United States) | Background job execution | Germany (EU), United States | DPA, SCCs, SOC 2 Type II |
Vercel, Inc. (United States) | Application hosting | Germany (EU) | DPA, ISO 27001, SOC 2 Type II |
AI model providers: each contractually undertakes not to use submitted content for training its models; retention is limited to a short period (abuse detection, zero to 90 days depending on the provider, extendable for flagged content), after which the content is deleted.
Provider | Processing location | Safeguards |
|---|---|---|
Anthropic, PBC (United States) | Worldwide | DPA, DPF, ISO 27001, SOC 2 Type II |
Marqo, Inc. (United States) | United States | SCCs |
Cerebras Systems, Inc. (United States) | Worldwide | SCCs, SOC 2 Type II |
Google, LLC (United States) | Worldwide | DPA, DPF, ISO 27001, SOC 2 Type II |
Liquid AI, Inc. (United States) | United States | SCCs, ISO 27001, SOC 2 Type II |
Mistral AI, SAS (France) | EU | DPA, ISO 27001, SOC 2 Type II |
Munito AI France, SAS (France) | EU | DPA |
OpenAI, LLC (United States) | United States | DPA, DPF, ISO 27001, SOC 2 Type II |
LlamaIndex, Inc. (United States) | United States | DPA, SCCs, SOC 2 Type II |
OpenRouter, Inc. (United States) | United States | SCCs, SOC 2 Type II |
Subquadratic Inc. (United States) | United States | DPA, SCCs |
TypeSafe AI, Inc. (United States) | Worldwide | DPA, SCCs, SOC 2 Type II |
Transfers outside the EU: where subprocessors are established in a third country, the transfer is governed by an adequacy decision (including the EU-US Data Privacy Framework) or by standard contractual clauses, supplemented by a transfer impact assessment. Details per subprocessor on the page referred to above.
7. Assistance to the Controller
Examino assists the Institution, insofar as possible and taking into account the nature of the processing:
in responding to requests to exercise data subject rights (access, rectification, erasure, restriction, objection, portability); requests received directly by Examino, including from the student area, are forwarded to the Institution without being answered on the merits, unless instructed otherwise;
in informing data subjects (Art. 13 and 14): Examino provides a template information notice for students and their legal guardians, which the Institution distributes;
with the obligations of Articles 32 to 36: security, breach notification (informing the Institution without undue delay and no later than 48 hours after becoming aware of it, with the information listed in Art. 33(3)), DPIA and prior consultation (Examino provides a pre-filled DPIA).
8. Fate of the data
At the end of the contract, or at the Institution's request, Examino deletes all papers and associated data (actual deletion, files included) or returns them in a structured format and then deletes them, unless a legal retention obligation applies.
Retention periods during performance: data is deleted by the teacher or the Institution's administrator from the application, or by Examino on documented instructions from the Institution (email to dpo@examino.ai). Examino recommends the following periods, which the Institution adapts in particular to the appeal periods applicable to it:
paper data, grades, comments, students' answers and work, student lists: current school year + 1 year (secondary) or end of the program + 1 year (higher education);
student accounts: deleted with the class or at the Institution's request; for access via the GAR, according to the GAR retention rules;
edit history of papers and gradings: kept with the paper and deleted with it;
authentication events: 12 months, then automatic deletion;
documents placed in the library trash: automatic permanent deletion after 30 days;
technical, operational and security logs held by infrastructure providers: 90 days maximum;
backups: deleted data is permanently purged as the rotation cycle proceeds, within 90 days at the latest.
Removal of a teacher from an institution workspace by an administrator of the Institution ends their access; the papers and data they processed on behalf of the Institution remain available to the Institution, which may request their deletion.
9. Automated decision-making
The service takes no decision about a student that produces legal effects or similarly significantly affects them based solely on automated processing (Art. 22 GDPR). Every grade and comment proposed by the AI is subject to review and validation by the teacher, who remains the sole decision-maker, and is only visible to the student after that validation. Scores calculated automatically for closed-ended questions in online assessments (multiple-choice quizzes, for example) result from rules set by the teacher, who can view and change them. The application logs the review of gradings (timestamped edit history).
10. Audit and documentation
Examino makes available to the Institution all documentation necessary to demonstrate compliance with Article 28 (Art. 30(2) record, security sheet, redaction technical sheet, list of subprocessors, model providers' no-training and retention terms, template student information notice) and allows audits, including by a mandated third party, with reasonable notice of 30 days, no more than once a year except in the event of a proven breach, at the Institution's expense and without access to other customers' data.
11. Term, governing law
This agreement takes effect upon the first processing of the Institution's student data (import of papers, creation of a class or student list, or first login of a student) and remains in force for as long as Examino processes data on its behalf. French law; jurisdiction of the administrative or judicial courts depending on the nature of the Institution.
For Examino
Mr. Dimitri Nicolas, President
For the Institution
Name, title, date, signature: