Privacy Policy

Last updated: 29 September 2026

Google Drive section added: 22 October 2026. Registration, approval, and quotas: 22 October 2026. Cataloging tracks and optional research: 22 October 2026.

This policy in Hebrew: מדיניות פרטיות.

This policy explains how MF Smart Research handles personal data on this website and in the MF Smart Research archive catalog. It is written in plain language. It is not a substitute for a signed agreement or for advice from your own lawyer.

Who we are

The vendor is MF Smart Research (Maaty Frenkelzon, מתי פרנקלזון), Warburg 4, Jerusalem 9654414, Israel.

Email: [email protected]

The service this policy covers

The MF Smart Research archive catalog is a universal archival cataloging system. An archive of any kind can use it. The service can include AI-assisted cataloging, Hebrew and Yiddish OCR, digitization support, and oral-history processing. Holocaust and Jewish collections are one possible domain. They are not the only one.

This policy covers:

  • the public website at mf-sr.com
  • the catalog and related processing we carry out for a customer
  • email and the website contact form

Roles

Two different roles apply, and it matters which one is which.

Archival content the customer provides. The customer, or the institution the customer represents, is the data controller for archival content and metadata it uploads or otherwise sends to us. That includes scans, photographs, films, audio, transcripts, catalog fields, and notes. MF Smart Research acts as a processor. We process that content on the customer's instructions to provide the service, as the agreement or order describes. If there is no written processing clause yet, this policy describes our practice, and we will put a written arrangement in place when the customer asks or when the law requires one.

Our own business contacts. For the public website, inquiries, and account and billing contacts that we administer ourselves, MF Smart Research decides the purposes and means and acts as controller.

Data we collect

Account and contact details. If you write to us or use the contact form, we receive the details you send. The form asks for name, institution, email, phone, and a message. You can reach us by email at the same address without using the form. If you open an account for the catalog, we process the details needed to run it, such as a name, an email address, the institution, and sign-in credentials or another sign-in method agreed for that account.

Usage and log data. Servers that deliver the website or the catalog may record technical logs, such as IP address, date and time, browser or device type, and the page or action requested. We use those logs to operate, secure, and troubleshoot the service.

Customer archival content and metadata. Customers send material so that we can catalog, read, or digitize it. That material, and the metadata around it, may include personal data about people named or shown in the collection, including people who are still living. We process it on the customer's behalf.

We do not buy marketing lists, and we do not sell personal data.

Why we use it

We use personal data to:

  • answer inquiries and provide the catalog and related services
  • run accounts, keep the service secure, and troubleshoot faults
  • produce transcripts, catalog drafts, and similar outputs for the customer who supplied the content
  • keep records of what was requested and what was agreed
  • meet a legal duty

The website contact form is delivered by email through FormSubmit (formsubmit.co) to [email protected]. The same inquiry — name, institution, email, phone, message, and the language of the page — is also sent to an MF Smart Research notification service at mondial.mf-sr.com so that we can see the inquiry.

AI, OCR, and model training

To provide the service we may run customer content through optical character recognition (OCR), handwriting recognition, and other AI features. That processing produces transcripts, catalog drafts, and similar outputs for that customer. A person should review those outputs before they are treated as a finished record.

Customer content is not used to train models for other customers unless the customer agrees in writing.

Sensitive and historical personal data

Archival collections often contain information about living people, and about people who have died but whose records still identify living relatives. That information can include what the GDPR calls special-category data, such as health, religion, ethnic origin, or political opinions. It can also include testimony, correspondence, photographs, and oral-history recordings.

Holocaust and Jewish collections are one domain where this arises. They are not the only one. The same care applies to any archive.

For content the customer provides, the customer chooses the purpose and must have a lawful basis. We process that content on the customer's instructions, to provide the cataloging, OCR, digitization, and oral-history features the customer asked for. We do not build our own marketing profiles from it.

How long we keep data

We keep personal data only for as long as we need it for the purpose we collected it for, and for any period the law requires.

  • Inquiries sent through the website or by email: for as long as we are corresponding with you, and afterwards for as long as we need a record of the request and of what we replied. We then delete them or reduce them to what we still need.
  • Account details: for the life of the account, and afterwards for the time needed to finish billing, security, and handover. Where an agreement sets that time, the agreement controls.
  • Customer archival content and metadata: for the term of the service. When the service ends, we delete it or return it as the agreement states. If the agreement does not set a time, we delete or return it within a reasonable time after you ask, except for copies the law requires us to keep and copies that remain in routine backups until those backups are overwritten in the ordinary course.

You may ask us to delete personal data earlier. We will do so unless we still need it to provide a service you asked for, or the law requires us to keep it. For archival content, we act on the controller's instruction.

Security

We use measures appropriate to the material. They include limiting access to people who need it for the service, and encrypted connections (HTTPS) on the public website. Where the agreement provides a dedicated server for an archive, that archive's content is not placed in a shared tenancy with other customers.

We do not claim a certification such as ISO 27001 or SOC 2.

Hosting and other services

The public website is hosted on GitHub Pages. Pages load typefaces from Google Fonts and icon styles from the Lineicons CDN. Your browser contacts those providers when it loads a page.

The catalog may be hosted on Google Cloud. The hosting location may be in Israel or in a region of the European Union.

Website inquiries also pass through FormSubmit and the notification service described above. A customer who has a processing agreement will be told, in the way that agreement describes, before we engage a new sub-processor to handle that customer's archival content.

International transfers

MF Smart Research is established in Israel. People who use the website or the catalog may be elsewhere. If personal data is transferred from the European Economic Area, the United Kingdom, or another place that restricts transfers, we do so only where the law allows that transfer, for example under an adequacy decision, standard contractual clauses, or another mechanism that applies. A customer that sends us archival content from outside Israel is responsible for having its own lawful basis to share that content with us.

Your rights

If the GDPR applies to you, you can ask to access, correct, or delete personal data, to restrict or object to certain processing, and to receive data you provided in a portable form. Where we rely on consent, you can withdraw it. You can also complain to a supervisory authority in your country. These rights are not absolute, and a request about archival content may need to go to the customer who is the controller. We will help that customer respond when we are the processor.

Under Israel's Privacy Protection Law, 5741-1981, you may ask to review personal data we hold about you and to correct it if it is not accurate, to the extent that law gives you that right. You may also contact the Israeli Privacy Protection Authority (הרשות להגנת הפרטיות).

To use any of these rights, email [email protected] with "Privacy" in the subject line. Please say which right you want to use and how we can reach you. We may need to confirm that the request comes from you.

Cookies and analytics

This website does not use analytics or advertising cookies. The pages do not load Google Analytics, an advertising pixel, or a similar tracker, and they do not ask you to accept optional analytics. The site search runs in your browser against a static index. It does not send your search as a tracked analytics event.

We do not set our own tracking cookies from the site's scripts. The host that serves the website may keep ordinary server logs. Loading Google Fonts and the Lineicons CDN sends those providers the usual technical data for a file request, such as an IP address. We do not control their logs.

The Drive catalog app at /app/ is the exception in the next section. It loads Google's sign-in script and the Google Picker. The other pages of this website do not.

Google Drive catalog app

The page at /app/ is how a signed-in Google user sends chosen files into the catalog.

What the app can access. The app asks Google for two non-sensitive Drive permissions: drive.file and drive.install. drive.file lets the app read and create only the files you pick in the Google file picker, the files you open with the app from Drive, and files the app itself creates. It does not let the app see the rest of your Drive. drive.install lets the app appear in Drive's "Open with" menu. The app does not request a restricted Drive permission, and it does not ask to read every file in your Drive.

Google account details. After you sign in, the app reads your Google account display name and email address so it can show who is signed in. Drive access uses a short-lived access token kept in the browser tab. A separate Google ID token tells our gateway which account is signed in. The app does not receive or store a Google password, and it does not keep a refresh token. The gateway checks the ID token's signature and does not store the token.

Registration and approval. Before a file can be sent, you register in the app. The form asks for your name, your organization or the words "private individual", whether the account is an institution or an individual, the purpose of the cataloging, and the languages of the material. We store that registration, together with the Google account identifier (sub) and the verified email address from the ID token, on the gateway at api.mf-sr.com (Cloudflare D1). The status is "pending approval" until Maaty Frenkelzon ([email protected]) approves or rejects it. A new registration sends an email to that address, and it also appears on the approval list in the app. That list is available only to the Google account [email protected]. The app does not offer the upload control until the account is approved. You can ask us to delete the registration by emailing [email protected] with "Privacy" in the subject line.

Quotas. An approved individual account may send 10 items in total, of any kind, with no time limit, unless we set a different number for that account. An approved institution on the pilot may send 100 documents, 10 photographs, and 5 films, with no time limit, unless we set different numbers. The gateway checks the remaining quota before it forwards a file. When the quota is used up, the app explains that and does not send the file. These are quantity limits, not a subscription period.

Do research. If you turn on “Do research” and choose a tier, the gateway sends the title, the date, and the people, places, organizations, and events named on the draft to the open sources configured for that tier. Those sources can include Wikidata, Wikipedia, VIAF, GeoNames, Europeana, the Internet Archive, and the Library of Congress. A source that needs a key is skipped when no key is set. The full transcription is not sent for this search. The app shows the result as a research note, separate from the catalog fields, with a short summary, findings, and a list of sources. A finding is added to the record only if you choose “Add to record”. Research uses its own quantity quota, and only the tiers enabled for the account. A pilot account can use the Basic tier. The price labels in the app are placeholders, not a charge by themselves. The gateway keeps a short-lived cache of source responses and a count of research use.

Files. When you pick a file or open one with the app, the browser downloads that file from Drive. It sends the file to https://api.mf-sr.com only after your registration is approved and only while quota remains. The gateway then forwards the file to the MF Smart Research catalog at films.mf-sr.com so the catalog can run OCR and draft a catalog record. The browser also sends the languages you chose: the interface language, the language of the catalog record, and the document language, or auto-detect for the document. The record is shown to you in the app. If you ask, the app saves the record as a new file in an app folder in your Drive named "MF Smart Research catalog", or in a folder you pick. The app does not change the original file. The browser never holds the Cloudflare Access service token used for that forward. The Drive access token is not sent to the gateway.

Limited use of Google user data. Google user data is used only to provide this cataloging feature: to identify the account for registration, approval, and quota, to read the files you selected or opened, to produce and show the draft record, and to save a file you asked to save. We do not sell Google user data. We do not use it for advertising. We do not use it to train models for other customers. We do not transfer it except to the catalog service that processes an approved file for you, to the open sources named above when you turn on Do research, and as the law requires. People at MF Smart Research do not read those Drive files unless you ask for help with a specific file, we need to look into a security incident, or the law requires it. The owner does see the registration fields in order to approve or reject the account.

How long. The access token and the ID token stay in the browser tab until you sign out or close the tab. The registration is kept while the account exists, and afterwards as needed to show what was approved and to secure the service. Retention of archival content on the catalog server is described above, under customer archival content. You can disconnect the app on the Google Account permissions page. Email [email protected] to ask for the registration to be deleted.

Signing in loads Google's own sign-in and file-picker scripts. Those scripts may set cookies Google uses to complete sign-in. They are not analytics or advertising cookies set by us.

Children

The website and the catalog are meant for institutions, researchers, and other adults. We do not knowingly collect personal data from children through the website. A collection may contain historical records about children. When a customer includes that material, the customer is responsible for the lawful basis, and we process it only as part of that customer's content.

Changes

We may update this policy. When we do, we will change the date at the top of this page. If a change affects how we process a customer's archival content, we will tell customers who have an active agreement, in the way the agreement describes or by email.

Contact

MF Smart Research (Maaty Frenkelzon, מתי פרנקלזון) Warburg 4, Jerusalem 9654414, Israel [email protected]

For a privacy request, put "Privacy" in the subject line. For a security concern, put "Security" in the subject line. See also the support page and the terms of service.

This policy is effective from 29 September 2026.