This policy explains how IAJUS.pt processes your personal data, in compliance with Regulation (EU) 2016/679, the General Data Protection Regulation (GDPR), and with Lei n.º 58/2019, de 8 de agosto, which ensures its implementation in the Portuguese legal system.
1. Controller
The controller of the data collected through iajus.pt is Celeris Juris Tecnologia e Inteligência Jurídica LTDA, a company incorporated under Brazilian law that operates the IAJUS brand. For any question on data protection, or to exercise your rights, contact:
Privacy email: privacidade@iajus.pt
We will reply to your request without undue delay and, in any case, within one month of receipt, under art. 12.º, n.º 3, RGPD.
1.1. Controller established outside the European Union
Celeris Juris Tecnologia e Inteligência Jurídica LTDA is established in Brazil, a third country in relation to the European Union. Because we offer the service to people located in the Union, this processing is subject to the GDPR (art. 3.º, n.º 2). You can send us any question or exercise your rights directly through privacidade@iajus.pt, and lodge a complaint with the CNPD under section 12.
2. How the service is designed: minimisation of personal data
IAJUS.pt is a legal research service. What it makes searchable today is case law published for public consultation, and it expects the query to be framed in legal theses and legal situations: the question of law, the legal concept, the claim, the cause of action. We do not need identifying data of third parties to answer you, we do not ask for it and we ask you not to include it in your queries. This is the concrete application of the principle of data minimisation (art. 5.º, n.º 1, al. c), RGPD) to the design of the product, and not merely a recommendation for use. The corresponding duty is in § 6 of the Terms (courtesy translation).
The searchable collection is made up of public information: decisions published for public consultation, with a link to the origin when available and the status of the source identified. We do not collect unpublished case files, filings or dossiers, and we do not feed the collection with what users search.
What this does not mean. Minimising is not the same as not processing personal data, and we would not tell you that we do not process it. We do process it: your account data, the record that you searched and when, and the technical access logs are your personal data, even when the question is an abstract thesis. And the text of the query itself is processed at the moment you search, because it is transmitted to be converted into a vector. Section 3.2 describes the processing and retention by channel. Published decisions may contain third parties' data despite the anonymisation practised by whoever published them: we do not promise you a collection free of them. The next section lists everything that is processed, with its basis and period.
3. What data we process
We distinguish three blocks, because their nature and their basis are different: your account data, what results from your queries, and the public collection, which is not collected from you.
3.1. Account, authentication and access data
When you are granted access to the service, we process the data needed to identify and authenticate you: name, email address and access credentials (stored encrypted or as a cryptographic hash, never in plain text), as well as the professional capacity and the organisation you tell us about. If a paid option exists, we also process the billing data that the law requires.
3.2. Your queries: what is recorded and what is not
The answer depends on the channel through which you search, and the boundary is this:
- Search through the website: the text of the question is processed to run the search, but it does not stay in the usage logs. These logs record the feature, the date and time, the structured filters and opaque identifiers of the results.
- Tools for querying the collection through the MCP connector or a plugin (for example, searching decisions and precedents from Claude or ChatGPT): the text of the question is processed to run the search. After automatic filtering, we keep a version of the term for 90 days to diagnose slow queries and improve search; expired terms are deleted daily. The filtering replaces CPF, CNPJ, OAB references, email and telephone number where it recognises them, but it may leave names and other identifiers. For that reason, the term is not anonymous.
- Lex search through the MCP connector or a plugin (standard search and deep search, in beta): here the conversation is kept. The text of your question, the context and constraints you give and Lex's answer are stored in a history specific to the MCP and plugin channel, separate from the Studio history, linked to your account and stored encrypted. The conversation can be resumed for 7 days from its creation, with no renewal, and the content remains stored, with no automatic deletion, until you delete it (section 9). The text of the question is also processed by artificial intelligence models to prepare the answer.
- Channels with conversation history — assisted conversation, deep research and analysis of a legal filing on the platform, WhatsApp and email: in these, the content of the conversation is kept to form your history, so that you can resume and consult it. That history is linked to your account, exists to serve you, does not feed advertising and is not visible to other users, and you can delete it at any time, all at once or channel by channel. The deletion is real — the content is deleted, not hidden.
Usage logs without content and filtered terms are stored separately, but can be linked by an internal identifier. That link may relate a query to the use of the service, so we do not claim that the term is anonymous. Access to filtered terms is restricted to authorised staff with an administrator profile. Conversation histories, including Lex's, remain separate and follow the rules described for each channel.
A consequence of particular interest to you. Anyone who practises as a lawyer is bound by professional secrecy (art. 92.º do Estatuto da Ordem dos Advogados [the Statute of the Bar Association]), and a legal research question can reveal the thesis, the strategy or the identity of a client. In the MCP or plugin tools for querying the collection, a filtered version of the term is kept for 90 days; the filtering may leave names and other elements that identify someone. For that reason, avoid including identifying data of clients. Restricted access and periodic deletion do not guarantee that all identifiers are removed, nor do they make the term anonymous. Where there is conversation history, including Lex search through MCP or a plugin, the history is yours, is not read by other users, does not feed advertising and can be deleted at your request, at any time, under section 9.
In search by meaning, the text of the query is sent, at the moment of the search, to the processor named in section 6 to be converted into a numerical vector; it may remain in the service's temporary memory for up to 24 hours. Separately, collection searches made through the MCP connector or plugins, including those started by Lex, may generate a filtered term kept for 90 days for diagnosis and search improvement (sections 3.2 and 8). This log keeps terms of collection searches, not the full conversation history. The transfer outside the European Economic Area is described in section 7.
The filtered terms kept for diagnosis correspond to collection searches made through the MCP connector or plugins, including searches started by Lex; they are not the full conversation history. Conversation history is not used to train or fine-tune the service, in line with the purpose already described for those channels. Lex's histories and those of the other channels remain subject to their own retention and deletion rules set out in section 3.2.
3.3. Technical access logs
The operation and security of the service generate logs: IP address, date and time, browser identification, requested address and result of the request. They are also personal data and serve to detect and stop abuse, fraud and unauthorised automated use, and to diagnose faults.
3.4. Public collection (not collected from you)
The decisions that the service searches are not supplied to us by you: they are obtained from publicly accessible sources or from historical mirrors identified by the service. They may contain third parties' personal data, even where whoever published them practises anonymisation. As to that content:
- we do not enrich it with data from other sources, nor do we try to identify whom the sources anonymised;
- we do not build profiles of parties, witnesses, judges or counsel, nor do we classify people;
- we keep the link to the origin, so that you can consult the page and its current version; the service identifies it as official only when it has verified it;
- if you are a data subject and believe that content in the collection affects you, write to us: we examine the request and, where you are right, we act on our own collection and refer you to the source that published the document, which is responsible for the original text.
3.5. Data you give us in the contact form
- Mandatory: email address and the message you send us.
- Optional: name, country, professional capacity (for example lawyer, solicitador/a [a regulated legal profession in Portugal], jurist, magistrate or law student), professional licence number (cédula profissional), organisation or firm, and telephone/WhatsApp number (which may have the dialling code of any country).
We do not ask you for special categories of data (art. 9.º RGPD) and we ask you not to include them in the free-text message.
3.6. Browsing data and audience measurement
Only with your prior consent do we collect usage data of the site through Google Analytics (via Google Tag Manager), for example pages visited, device type, origin of the visit and a cookie identifier. Until you give consent in the cookie notice, no analytics cookie is installed and no data is sent to Google.
4. Purposes
- Providing the research service: authenticating your access, running the queries you make and returning the results linked to the source.
- Recording the use of the service without its content — which feature, when and with what result — to grant you access, give you support when you ask for it and size the service.
- Using filtered terms of collection queries via MCP or plugin to diagnose slow queries and improve search.
- Keeping the history of your conversations, in the channels that offer it (section 3.2), so that you can resume and consult them — and delete it whenever you want.
- Replying to your contact and providing the information requested.
- Informing you about access to the service, when you have asked for it.
- Complying with legal obligations to which we are subject, in particular billing and accounting retention, where applicable.
- Measuring audience on the site in order to improve it (only with consent).
- Ensuring the security of the service and preventing abuse and fraud.
We do not use your data for third-party advertising and we do not sell it.
What we do not do. We do not profile the people mentioned in the collection, nor those who may be mentioned in your queries. We do not take automated individual decisions about people, within the meaning of art. 22.º RGPD: the service returns documents and summaries for human assessment, and the professional decision is always that of whoever uses it. We do not use your data to assess your creditworthiness, your professional performance or the merits of any case.
5. Legal basis
| Processing | Legal basis (GDPR) |
|---|---|
| Providing the research service: account, authentication and running the queries | Performance of a contract: art. 6.º, n.º 1, al. b) |
| Sending the text of the query to the processor that converts it into a vector, so that search by meaning works (section 6) | Performance of a contract: art. 6.º, n.º 1, al. b). It is the operation that produces the result you asked for: without it, search by meaning cannot take place. Keeping a filtered version of the term for diagnosis is a separate operation, described in sections 3.2 and 8. |
| Keeping conversation history, in the channels that offer it (section 3.2) | Performance of a contract: art. 6.º, n.º 1, al. b). It is the feature provided to you — being able to resume and consult your conversations — and it remains under your control: you can delete it at any time, all at once or channel by channel (section 9) |
| Usage log without content (account, moment, feature, number of results) | Legitimate interest: art. 6.º, n.º 1, al. f). We weighed our interest in granting access, supporting those who use the service and sizing it against the reasonable expectation of the person making the query, and concluded that our interest prevails precisely because this log does not include the content of the query and involves no profiling or assessment of people. This balancing applies only to this usage log without content; it does not cover the filtered terms described in section 3.2. Conversation history is also not part of this log. You may object under art. 21.º, n.º 1 |
| Building and maintaining the searchable public collection (documents published for public consultation) | Legitimate interest: art. 6.º, n.º 1, al. f), to make search available over information whose publicity is imposed by law, without enrichment or profiling |
| Technical access logs, security and abuse prevention | Legitimate interest: art. 6.º, n.º 1, al. f) |
| Billing and keeping supporting documents, where applicable | Legal obligation: art. 6.º, n.º 1, al. c) |
| Contact form and communications about access | Consent: art. 6.º, n.º 1, al. a) |
| Analytics cookies (Google Analytics) | Consent: art. 6.º, n.º 1, al. a) |
| Security, abuse prevention and operation of the service, including diagnosis of slow queries and search improvement with filtered terms (section 3.2) | Legitimate interest: art. 6.º, n.º 1, al. f) |
You may withdraw consent at any time, without affecting the lawfulness of processing carried out until then (art. 7.º, n.º 3, RGPD).
6. Recipients and processors
To provide the service we use processors that process data on our behalf, under data processing agreements (art. 28.º RGPD):
- Google Cloud (Google Ireland Ltd. / Google LLC): hosting and infrastructure, including storage of filtered terms of collection queries.
- OpenAI (United States): converts the text of the query you write into a numerical vector, so that the search can look by meaning and not just for matching words. The sending takes place at the moment you search and is intended exclusively for that conversion: the text is not used to answer you by any other route, nor to identify you.
- Resend: sending and receiving the form's email messages.
- Google Analytics / Google Tag Manager: audience measurement, only with consent.
There is no sale or sharing of data for third-party marketing purposes.
7. International transfers
The controller is based in Brazil, a country that is not the subject of a European Commission adequacy decision. The transfer of your data to Brazil relies on the derogations of art. 49.º: the necessity of performing the service you requested and your consent.
This disclosure covers all the operations described in section 3, and not only the contact form: your account data, the text of the query while it is being processed, the filtered terms kept in the MCP or plugin channels, the conversation history in the channels that keep it and the technical access logs are processed under the responsibility of the entity identified in section 1, on the same transfer basis. In channels without history, the text is not kept as a conversation; in collection searches via MCP or plugin, a filtered version of the term is kept for 90 days (sections 3.2 and 8). The text is processed at the moment you search, and that is what this section covers. In channels with history, what is kept is the history already described in section 3.2, which remains subject to your deletion order, with no automatic deadline. It is one more reason why the service is designed not to need third parties' personal data in queries: the less data there is, the smaller the reach of any transfer.
Some processors may also process data outside the European Economic Area (for example, in the United States). Those transfers rely, like the previous one, on the derogations of art. 49.º — the necessity of performing the service you requested and your consent. We do not invoke here any adequacy decision, Standard Contractual Clauses or membership of a data protection framework: we will state one of those mechanisms only when we can document it, and this section will then be updated to indicate the mechanism that applies to each processor.
8. Retention
We keep each category of data only for as long as needed for the purpose that justifies it (art. 5.º, n.º 1, al. e), RGPD):
- Contact form data: for as long as the purpose lasts (replying to the contact and, if requested, communicating about access) and up to 24 months after the last contact, or until you withdraw consent, whichever occurs first.
- Account data: while the account remains active. If the account is closed, the data is deleted or anonymised, except what we must keep to comply with a legal obligation or to exercise or defend a right in proceedings.
- Usage logs without content: for the period needed for granting access, support and sizing the service.
- Filtered terms of collection queries via MCP or plugin: kept for 90 days; expired terms are deleted daily. You can ask for early deletion at privacidade@iajus.pt. The filtering may leave names and other identifiers, as explained in section 3.2.
- Conversation history, in the channels that keep it (section 3.2): for as long as you keep it. We have not set an automatic deletion deadline: the content stays until you delete it, and you can do so at any time, all at once or channel by channel. In Lex search via MCP or a plugin, the conversation can be resumed only for 7 days from its creation, with no renewal; the content remains stored, encrypted, until you delete it. The encrypted technical copies linked to the quota and to safe retry of the request are deleted in the automatic maintenance following the end of those 7 days or the deletion of the conversation.
- Technical access logs: for the period needed for the security and diagnostic purposes that justify them, after which they are deleted or aggregated so as not to allow identification.
- Billing documents, where they exist: for the period imposed by the applicable tax and accounting law.
- Audience analytics data: for the period defined in the Google Analytics configuration.
Once the periods end, data is deleted or anonymised according to its category. You can ask us at any time to delete your account, the usage logs linked to it and the filtered terms of collection queries.
9. Your rights
As a data subject, you have the following rights:
- Access to your data (art. 15.º).
- Rectification of inaccurate data (art. 16.º).
- Erasure (“right to be forgotten”) (art. 17.º).
- Restriction of processing (art. 18.º).
- Portability of data (art. 20.º).
- Objection to processing based on legitimate interest (art. 21.º, n.º 1). This right expressly covers the usage logs without content and the filtered terms of collection queries described in section 3.2, as well as the technical access logs of section 3.3.
- Withdraw consent at any time (art. 7.º, n.º 3).
- Not be subject to automated decisions that significantly affect you (art. 22.º). As explained in section 4, we do not take such decisions.
- Lodge a complaint with the supervisory authority (art. 77.º), under section 12.
To exercise any of these rights, write to privacidade@iajus.pt. Deletion of the conversation history (section 3.2) can be requested by the same route at any time, all at once or channel by channel, and is carried out as real deletion of the content, never as hiding. Lex search conversations via MCP or a plugin are deleted when you delete the whole history through Studio, with all channels at once, or when you delete your account; deleting only one channel does not reach them. You can also ask, at the same address, for early deletion of the filtered terms of collection queries described in sections 3.2 and 8.
10. Security
We adopt appropriate technical and organisational measures (art. 32.º RGPD): connections encrypted by HTTPS/TLS, minimisation of the data collected, credentials stored as a cryptographic hash, segregation of databases by purpose, role-based access control and logging of operations that access user data. The form and the service include mechanisms to prevent automated abuse. If a personal data breach occurs that implies a high risk to your rights, we notify you and notify the CNPD under arts. 33.º e 34.º RGPD.
11. Cookies
We use only strictly necessary cookies for the site to work (which do not require consent) and, with your consent, analytics cookies. You can accept or refuse analytics cookies in the notice shown on the first visit and change your choice by clearing the site's data in your browser.
12. Supervisory authority
If you consider that the processing of your data infringes the GDPR, you have the right to lodge a complaint with the competent supervisory authority in Portugal, the National Data Protection Commission (Comissão Nacional de Proteção de Dados, CNPD), www.cnpd.pt (opens in a new window) (art. 77.º RGPD).
13. Changes to this policy
This revision corrects the description of the existing retention of filtered terms of collection queries; it does not announce new collection. This policy may be updated. The version in force is always available on this page, with the revision date shown at the top.