Legal
Privacy Policy
Service: Argentine Legal Corpus via MCP (https://mcp.bravo47.com).
Last updated: November 6, 2026. · Versión en español
Your queries are not stored
What you ask is processed in memory and discarded. This is what does and does not remain on the server:
We do not store
- The text of your queriesProcessed in memory to answer, then discarded.
- What we answeredNeither the result nor the document you saw.
- Who you areNo accounts, sign-up or passwords.
- Traces in your browserNo cookies and no third-party analytics.
We store
- An anonymous usage countFor each call: which tool, the channel, whether it succeeded, how long it took and the error type. No IP and no identifier.
- Technical server logsIP address, date and time, path and user-agent, for security, rate limiting and an aggregate count of distinct addresses published on the home page (never any individual address). Kept with no fixed deletion deadline, for security reasons.
What you type into Claude, ChatGPT or another assistant is handled by that company under its own policy: we only receive the request to our tools. Full details are below.
What this service is
A read-only query server over official public documents (national legislation from InfoLEG, case law of the Supreme Court — CSJN — and of the Supreme Court of the Province of Buenos Aires — JUBA —, treaties published by the Foreign Ministry, provincial constitutions and codes from SAIJ, the national and the Buenos Aires Province official gazettes, and court directories). It requires no registration, has no user accounts, and does not collect personal information from the people who use it. The documents are public and are returned as the official source publishes them. It is offered at a single address, https://mcp.bravo47.com/mcp; the numbered addresses of earlier versions are aliases of it.
Data we process
- Search queries: the text of each query is processed in memory solely to execute the search and return results. We do not store query contents, nor share them with third parties. Semantic similarity (embeddings) is computed entirely on our own server — queries are never sent to external inference services.
- Access logs: we log IP address, timestamp, requested path and user-agent for security, diagnostics and rate limiting. We also use them to publish one aggregate usage indicator on the home page: how many distinct IP addresses received a response from the service. It is a single total computed on the server; no individual address is published or shared. These logs are kept with no fixed deletion deadline, for information-security reasons: they allow us to detect abuse and intrusions and to reconstruct incidents that are discovered later. They are used only for those purposes and are not shared with third parties.
- Usage measurement: to learn which tools are used and how fast they respond, we record for each call the tool name, the channel, whether it succeeded or failed, how long it took and, on failure, the error type. We do not record who made the call (no IP, no identifier), nor the text of the query or its result.
What we do not do
- No cookies or tracking technologies.
- No third-party analytics. Fonts are served from our own server — visiting these pages triggers no third-party requests.
- We do not sell, rent or share data with third parties.
- We do not build user profiles.
Infrastructure processors
The service is hosted on a private server provided by Contabo GmbH. Traffic is encrypted with TLS (Let's Encrypt certificates). No other provider processes queries.
Your rights
You may request information about, or deletion of, access-log data associated with your IP address by writing to the contact below. Since we keep no accounts or persistent identifiers, access logs are the only data linkable to a person.
Changes to this policy
Changes will be published on this page, updating the date in the header.
Contact
gbaros@palermo.edu — Law School, Universidad de Palermo (Palermo E-Law / Innovation Hub), Buenos Aires, Argentina.