Privacy Policy
Last updated 9 November 2026. The data controller is Jorge de Diego, an individual who provides this service under the name “Rumbo”. Contact: soporte@hellorumbo.com.
What we collect
- Account data — your email address and your name, as you give them to us or, if you sign in with your Google account, as Google gives them to us, together with the link to your Google profile picture and the identifier of that account.
- Profile data — your profession, seniority, self-reported AI experience, the tools you use and your daily goal. We use this to choose which missions you get.
- Learning data — the work you submit for each mission, your answers in practice, the messages you send to the AI coach, your scores (your AI Score included), skill estimates, XP, streak and achievements; and, where Learn's «Saved» is open to you, the tricks and missions you save and the names you give your folders, which only you can see.
- Work data — your Work conversations, attached documents and the text read from them, the deliverables Work makes for you (documents, presentations, spreadsheets, charts and dashboards) with each of their versions, selected models and response feedback.
- Work profile — if you fill it in: your role, responsibilities, recurring tasks, the tools you use, what takes you time, what you may not share with an AI, a note, and whether you let ZEN use it. You can see, change and delete it in Work.
- AI kit and ideas — the templates and processes you keep in your AI kit, and what you do with the ideas Work suggests for your job: whether you open, start or put one aside, the steps you take and whether you say it was useful.
- What Work’s capabilities being tested add — only if you have them (see “Work capabilities being tested”): the notes ZEN keeps about you; the images and files Work makes for you; the pages an answer cites, the code it ran and what that code showed; the steps of an answer and, while it goes on in the background, what it has written so far; and the outline and notes of a long document.
- Usage data — which screens you open, which missions you start and finish, which language you use Rumbo in, and whether you see the phone, tablet or computer layout (from the width of your browser window, nothing else about your device), so we can tell whether the product is working. If you have any of Work’s capabilities being tested, we also record, for each answer, how many searches, sources, code runs, files, images, notes kept or forgotten, sections and steps it had and how long it took — never a word of what you wrote or of the answer — and some of what happens on the screen: a notice shown or clicked, notices switched on, an answer followed again, a note undone, a long document asked to go on. Nothing is kept in your browser to measure usage. What your browser reports while you are signed out we store with nothing that identifies you or links one visit to another, and we use it only in aggregate. While you are signed in it is linked to your account, and our servers group it into visits: a visit ends after 30 minutes without activity. Until October 2026 your browser kept a random identifier and sent it with what it reported; it no longer keeps one, and erases the one it had the next time you open Rumbo, but what it reported then still carries it, and so does an account created with an email address before then, until the account is deleted (see “Deleting your account”).
AI processing
To give you feedback, we send the following to our AI provider: the mission you are working on, the files it contains, the messages you send the coach and the work you submit. These are processed to produce coaching replies and an evaluation, and are returned to you. Do not paste real confidential company data into a mission. All missions use fictional companies and entirely synthetic financial data.
In Work, we send your conversation context, included files, profession and linked learning objective to the provider of the model that answers — the one you chose, or the one Rumbo picks when you leave it on Auto — to generate a response. Only attach material you are authorised to share. Provider processing and retention depend on the provider and the configured account. Deleting a conversation removes its messages and files from RUMBO; it does not undo processing already performed by a provider.
A provider may keep the beginning of a request for a few minutes, so that the next request that begins the same way is answered faster and at lower cost. When ZEN can use tools in Work and the model that answers is Anthropic’s, Rumbo asks Anthropic to do so: the beginning of the request — the tools’ descriptions and your conversation so far — is kept for 5 minutes after it was last used, and is only ever reused by a request of Rumbo’s that begins exactly the same way.
Work can read the text of a Word or Excel file you attach on our servers. That text is kept with the file and deleted with it, and it is sent, as part of the file, to the provider of the model that answers: OpenAI or Anthropic. When you ask Work for a deliverable, what the model wrote for it is kept so you can see, download and change it, and the deliverable you have open, or the last one changed, goes with each message of that conversation to the provider of the model that answers.
If you build your Work profile by talking with ZEN, your answers go, like any Work message, to the provider of the model that answers. After each answer, the conversation is also sent, without your name or email address and at most 12,000 characters, to the provider we use for Learn, today Anthropic, to prepare as you go a draft that is not kept until you review it. Only if you switch on “Let ZEN use this profile” does the profile (about 1,500 characters at most), with the list of what you marked you will not share, go with each of your Work messages to the provider of the model that answers. You can switch it off or delete the profile at any time; it stops being sent from your next message. Deleting the profile does not delete the diagnosis conversation, which you can delete in Work like any other.
We record which model was used and how many tokens it consumed, so we can manage cost. We do not use your submissions to train models.
Work capabilities being tested
While we test them, only some people have these capabilities of Work; if you have none of them, nothing in this section applies to you. You will see each one you have in Work, and what “AI processing” says holds for them too: each provider’s processing and retention depend on the provider and the configured account.
Web search. The model that answers can search the internet and read pages while it answers, with its provider’s own search: OpenAI’s for OpenAI’s models, Anthropic’s for Anthropic’s. What it searches for is written by the model from your conversation, and goes to that provider and, through it, to the search service the provider uses. The model is told to search for the general fact you need and to leave your private data out of a search, but nothing else filters what it writes. If you use Rumbo in Spanish, the search is made as if from Spain, the country alone; we do not send your location. The pages are read by the provider: your browser does not connect to them unless you open one of the sources, which are kept with the answer.
Running code. The model that answers can write Python and run it in its provider’s own environment for code (Anthropic’s code execution, OpenAI’s Code Interpreter), to work with your data, calculate and draw charts. For that, the CSV and Excel files of the conversation that the model receives as they are go into that environment; with an OpenAI model, OpenAI also copies there the other files it receives with the conversation. When the answer is written we delete from the provider the copies we uploaded; in case that deletion fails, each carries an expiry of one hour. Whatever else is left in that environment, the provider keeps and deletes under its own rules. The files the code writes, such as a chart or a workbook, are brought back to your private Work storage, and the code and what it showed are kept with the answer.
Images. Work can make an image from a description, or change one you attached or one it made, with OpenAI’s image model, whichever model answers. The description, which the model writes from what you ask, and the image to change go to OpenAI, with nothing else of yours: not your name, and no identifier. The image is kept in your private Work storage, with its description, until you delete the conversation or your account.
ZEN’s memory. Only if you switch on “Let ZEN remember” under “What ZEN remembers” — it is off until you do — ZEN keeps short notes about you from what you tell it about yourself, your work and how you like things done — at most 100, of up to 300 characters each — and shows each one it keeps under the answer, with “Undo”. The notes, at most about 8,000 characters of them, go with each of your Work messages to the provider of the model that answers. ZEN is told never to keep anything about health, sex life or orientation, religion or beliefs, political opinions, union membership, ethnic origin, or genetic, biometric or criminal data, nor, where ZEN uses your Work profile, anything of the kinds you marked in it that you will not share, nor a password, a key or a card or account number; a filter refuses a note that reads as one of those, but a filter of words can let one through. Under “What ZEN remembers” you can read, change and delete any note, or all of them, and add your own; a note you delete is no longer sent from your next message. Deleting a note does not take it out of the conversation where ZEN kept it, which still shows it until you delete that conversation; deleting that conversation does not delete the note. If you switch it off, ZEN stops keeping notes and sending the ones there are, which stay there until you delete them. The notes, and your answer to “Let ZEN remember” with its date, are deleted with your account and are included in “Download my data”.
Projects. You can group your Work conversations in projects. A project keeps the name you give it, the instructions you write for ZEN — at most 4,000 characters — and up to 10 reference files — at most 5 MB in all —, in your private Work storage. The instructions, the project’s name and its files go with every message of every conversation of the project to the provider of the model that answers that message, as the files you attach to a conversation do; the text read from a Word, Excel or PowerPoint file is kept beside the file. You can change the instructions, remove a file, take a conversation out of the project or delete the project at any time: what you change or remove is no longer sent from your next message. Deleting a project deletes its instructions and its files, but not its conversations, which go to “Recents”. Projects are deleted with your account and included in “Download my data” (the name, the instructions and the list of files; the files themselves are not).
Long documents. A document of many sections is written section by section, in further calls to the same model, each with what you wrote in the conversation; where you attached files, one of those calls reads them to note what each section should draw on. If the document does not fit in one answer, our server goes on with it in up to three further answers in the same conversation, which it starts by itself or when you choose “Continue now”. The document’s outline and those notes, which may quote your files, are kept with the deliverable and deleted with it, with the conversation or with your account.
Dictation. If you press “Dictate”, Work’s microphone, your browser asks your permission to use it and turns your voice into text with its maker’s speech recognition service (Google in Chrome, Apple in Safari, Microsoft in Edge), under that maker’s terms. Rumbo neither receives nor keeps the audio: the text appears in the message box and only reaches us if you send it, like any other message.
My style. If you save “My style” in Work, we keep only the colours, the typeface, the sizes and whether you want it by default, to apply it to what ZEN makes. The images you choose as a reference are read in your browser: they are neither sent to Rumbo nor kept. You can delete your style from “My style”; it is also deleted with your account.
Pages. ZEN can make a web page as a deliverable; it is kept like any other deliverable. If the page uses a library from cdnjs (Cloudflare) or jsDelivr, your browser downloads it from that service when you open it, and that service sees your IP address and which library is asked for, never what the page says.
The answer’s steps. While ZEN answers, you can see how long it has been working and each step it takes. When it is done, the steps stay folded under the answer until you delete the conversation; a step can say what was searched for, which site a page was on, a section’s title or a file’s name.
Answers that go on in the background. An answer can keep going on our servers after you close the tab, and you can follow it again from any browser where you are signed in. While it goes on, we keep what it has written so far and each of its steps, with what they show (what it searched for, the code it ran, a note it kept), so that it can be followed again. What it has written is cleared from there as soon as the answer is stored in the conversation. The steps are removed after a day, and sooner if you delete the conversation or your account.
To tell you that one of those answers has ended, Rumbo shows a notice on the screen. If you choose “Tell me when it is done”, your browser asks you whether Rumbo may show you its notifications, which show the conversation’s title. So as not to tell you of the same answer twice, your browser remembers for a day which answers it has told you of (their identifiers only), and it also remembers whether you said yes and that you have this capability.
What ZEN learns from your work (being tested)
While we test it, only some people have this, and it starts only if you say yes in “What ZEN has learnt about you”; if you have not, nothing in this section applies to you.
What is read. All your Work conversations, and on each pass only the new messages. Of each conversation we send, without your name or e-mail address, at most 6,000 characters of what you wrote and short excerpts of ZEN’s answers (at most 300 characters each). Of a file we send only its kind (for example, “xlsx”), never its name or contents.
Who processes it. The provider we use for Learn, today Anthropic, sums up each conversation as data: the tasks you do in it, by hand or with AI, how often, the tools you name and how you use AI. Our server checks each summary against your own words and drops what it cannot find in them and what looks like health, beliefs, political opinions, trade union membership, ethnic origin, sex life, genetic, biometric or criminal data, or a password.
What is kept. One summary per conversation, and a portrait our server builds, without AI, from those summaries, from what you did in Learn (tricks, missions and practice), from your Work profile if you let ZEN use it, and, if you switched on “Let ZEN remember”, from how many notes ZEN keeps about you. We also keep what we suggested to you, why, and what you did with each suggestion.
What for. To suggest tricks, missions, practice and ideas from Rumbo’s catalogue, each with its reason and the conversations it rests on; to rank Learn by what you do in Work; for Learn’s tutor to use your real tasks and ZEN’s notes as examples; and for ZEN, in Work, to know what you have learnt. To choose the suggestions, your portrait and a list of options from the catalogue go to the same Learn provider. Learn’s tutor, from the same provider, receives with your messages up to five of your most frequent tasks and, if you switched on “Let ZEN remember”, up to ten of the notes ZEN keeps about you, as many as fit; ZEN, in Work, receives with yours only the titles of the tricks you did lately and the skills you find hardest.
Your control. In “What ZEN has learnt about you” you can see everything we have learnt, hide a task, a skill, a tool or a whole conversation (which is not read again), pause it and delete it all. Deleting it removes the summaries, the portrait and the suggestions, and withdraws your consent. Deleting a conversation deletes its summary.
Tools you connect to Work
In Work you can connect tools you use, such as Gmail, Google Calendar, Google Drive, Google Docs, Google Sheets, Google Chat, Notion or Linear, so that ZEN can look things up and act in them for you. You choose which tools to connect and, action by action, whether ZEN may use it without asking, must ask you every time, or may not use it. Anything that writes, sends or deletes asks you first unless you choose otherwise, and an action the tool marks as destructive, such as sending an email, always asks, except in an agent you set to «Do everything without asking»: that agent does what its task needs with the tools you gave it without asking you, sending email included. When an agent works on its own, we email your address when it finishes, fails or needs you to decide something; the email says only the agent's name and what happened, never its work, and you can turn it off in the agent.
You connect a tool by signing in on its own page, where you approve what Rumbo may do; the tool then gives Rumbo a token to act for you within what you approved. We keep those tokens encrypted, and they never reach your browser, an AI model or our logs. When ZEN uses a tool, what it asks goes to that tool, and what the tool answers goes, as part of your conversation, to the provider of the model that answers.
Under each answer, and kept with the conversation, is a record of every action ZEN took: the action; what it asked (at most 12,000 characters) and, where the action changed something, what the tool answered, such as the id and link of what was made (at most 800 characters); a description of what it did, in your language, which can name what it searched for, the sender or recipient it searched by, dates, counts, the subject, title or name of the email, event, file, document, Chat space, cells or other item it read or changed, and, for any other tool, the kind of thing it acted on (a page, an issue, a task…) and how many people it named; and a readable summary of what came back, of at most 400 characters, such as the subjects and senders of the emails it found or the first lines of a document. So that ZEN does not repeat what was done and can go on from it, part of that record of the conversation's latest actions (the action, an excerpt of what it asked and of what came back) goes again, with each message, to the provider of the model that answers, along with your browser's time zone, to know what day it is for you.
When ZEN asks your permission before an action, the request is kept as it was shown to you: what the action would do — for an event, its title, its times, its place and description where the action changes them, its organiser, up to 3 of its guests by name and address, and every guest it adds or removes; for a spreadsheet, the current and the new values of the cells it would write (at most 36 cells of up to 80 characters each); for a Chat message, its text, the name of the space or conversation it goes to and the names of a few of its members; for a Drive file, its name, its owner’s address and, for a new one, its text; for a document, its title and the text the action adds or replaces; for any other tool, each value the action would send, under the label the tool gives it, with the people it reaches and where it would land shown first, the tool’s own name and what the tool says of itself (that it only reads, or can delete or change things) — and where the idea came from: the kind of thing ZEN had read before asking (an email, an invitation, a file, or a page or an issue of another tool…), who sent, organised or owns it, by address, for the Google tools Rumbo runs itself, and the names of the connections it was read from.
When Rumbo shows you, as cards under an answer, items ZEN read for that answer (events, emails, files, documents or Chat messages, and items another tool returned with a title, such as pages, issues or tasks) — the ones ZEN chooses or, where it chooses none, up to 8 that Rumbo picks by kind — each card keeps what it shows: the item’s title or subject, who sent or owns it, its date, time and place, how many people are on it and the initials of up to 3 of them, whether it is unread, important or shared, the first words of an email or a message (at most 160 characters), for another tool’s item its short reference, its status and the address of the site its link opens on (only the tool’s own web addresses), and its link, with ZEN’s headings, notes and suggested next steps where ZEN chose the cards. The cards are kept with the answer.
Several actions at once, and undoing. When ZEN asks your permission for several actions at once, each request also keeps its place among them, so that they are shown to you and decided together. When ZEN changes something in one of your Google services (an event it creates, changes or answers, a draft or a file it makes, cells it writes, a Chat space it marks as read), the record of that action also keeps what undoing it needs, within 12 KB: the identifier and version of what it made or changed and, for a change, what was there before as ZEN read it (an event’s title, times, place, description and guests by address, your earlier answer and its comment, the cells’ earlier values, when you had last read the space); and, if you undo it, when and how that went. You can undo it for 15 minutes. What undoing needs is deleted once it is used; if it is not used, it is deleted the first time, 20 minutes or more after the action, that you send a message in Work or ZEN changes something for you. When and how an undo went stays with the action’s record.
All of this is kept with its conversation and deleted when you delete the conversation or your account; disconnecting a tool does not delete it.
The tools you connect are your own providers, under your own agreements with them: Rumbo sends them only what ZEN asks on your behalf. You can disconnect a tool at any time in Work: its tokens are deleted at once and, where the tool allows it, it is asked to revoke Rumbo’s access. You can also remove that access from the tool itself; for Google, in your Google Account, under third-party apps with access to your account.
Google user data. Rumbo’s use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements. We use what we receive from your Google account only to do what you ask of ZEN in Work and to show it to you. We do not sell it, use it for advertising, or let anyone read it except at your request, to keep the service secure, or where the law requires it. We do not use it, and the AI providers we send it to do not use it, to develop, improve or train generalised AI or machine-learning models.
Files you choose from Google Drive. When you use “Choose from Drive”, Google’s own file picker opens in your browser, with a short-lived permission Google gives your browser for that alone; it is never sent to Rumbo or kept. The files you choose, and only those, can then be read by Rumbo’s Google connections: Google remembers that you gave them to Rumbo until you remove Rumbo’s access in your Google Account, and disconnecting a tool in Work does not take them back. With your message Rumbo keeps each file’s name, type and Google identifier (at most 10 files a message), so that you see them and ZEN can read them when it answers; they are deleted when you delete the conversation or your account. Where the button is offered, your browser loads the picker’s code from Google when you open Work.
Using Rumbo from other apps
You can add Rumbo to an AI app you use, such as Claude (by Anthropic) or ChatGPT (by OpenAI), as a connector. The app sends you to Rumbo to sign in, and Rumbo’s own page asks what that app may do: search and recommend the tricks of your department and see which ones you have done; read your Work conversations, deliverables, projects and AI kit, if you tick it; read what ZEN knows about your work (your role, tasks and tools, and what you do not want shared), if you tick it and only where you have already let ZEN use your profile in Rumbo; and save things in your Rumbo when you ask it to — a document as a new Work deliverable, a request in your AI kit, or whether a trick worked for you — if you tick it. Reading ZEN’s profile and saving start unticked. It can never send, delete or spend anything in Rumbo; what it saves is new, except your rating of a trick, which replaces your earlier one. A trick’s ready-to-use request is shared only for tricks you have already done in Rumbo.
When you then ask the app something that needs Rumbo, it asks Rumbo, and Rumbo answers with what you allowed: the tricks and, if you allowed it, the text of the Work conversations, deliverables, projects, AI kit items and profile it asks for, within the history your plan shows you in Rumbo. What it saves for you is kept in your Rumbo like anything else of yours there, and deleted the same way. What Rumbo sends goes to that app and its provider under your own account with them and their terms: for that use they are not our processors, and what they do with it is governed by your agreement with them, not by this policy. Nothing is sent to an app you have not connected, and nothing is sent unless the app asks for it while you use it.
For each app you connect Rumbo keeps its name, the address it sends you back to, what you allowed, when you connected it, when it last used Rumbo and, if you did, when you disconnected it; and, to recognise the app, short-lived codes and tokens, stored only as a one-way fingerprint (a hash) from which the token cannot be recovered. Access tokens last an hour and renewal tokens 30 days from their last use. In Account › Claude and ChatGPT you can see these apps and disconnect any of them, which stops it at once; deleting or deactivating your account does the same. The list of apps you connected is part of “Download my data”.
Legal basis
We process your account, profile, learning and Work data — what Work’s capabilities being tested add included, except ZEN’s memory — send our AI providers what is described above, and exchange with the tools you connect what is described under “Tools you connect to Work”, and answer the apps you connect Rumbo to what is described under “Using Rumbo from other apps”, to perform our contract with you: it is the service you signed up for.
There are two exceptions, both based on your consent. The first is your Work profile going with your Work messages, or to an app you connect Rumbo to: we send it only if you consent by switching on “Let ZEN use this profile” (and, for an app, only if you also tick it when you connect it). The second is ZEN’s memory: we keep and send your notes only if you consent by switching on “Let ZEN remember”. You can withdraw either at any time by switching it off.
What ZEN learns from your work (see “What ZEN learns from your work”) is based on your consent too: we process it only if you consent in “What ZEN has learnt about you”, and you withdraw it there with “Delete everything it knows”; while it is paused, nothing new is read and nothing is suggested.
We process usage data, record the cost of each AI request, keep error reports, technical logs and backups, and read the comments you send us on the basis of our legitimate interest in understanding, improving and protecting the product and in being able to recover it.
If you joined through a company invitation, we show that company what is described under “What your employer can see” on the basis of our legitimate interest, and the company’s, in knowing how the team it invited is taking part. The company does not decide with us how we process your data, and we do not process your data on its behalf.
What your employer can see
If you joined through a company invitation, an administrator or manager at that company can see who is in the team — each member’s name, email address, profession and role in the team, and whether they have finished onboarding — and the invitations still pending. They can also see aggregate progress for their team: how many people have onboarded, have started and finished a first mission and have been active in the last seven days (practice counts as activity), and, only once at least three people are behind each figure, how many missions have been completed, how many people are certified, average scores and average skill levels across the group, and what the team reports about using it in their work. A figure is also withheld while it could be worked out from another by subtraction, so an individual cannot be identified that way.
They cannot see your individual submissions, your coach conversations or your individual scores. Your Work conversations, files and deliverables, your Work profile (and whether you let ZEN use it), your Work projects, what ZEN remembers about you, what Work makes for you and your AI kit are also private to your account: nothing from Work is in the team’s figures. So is what ZEN learns from your work and what it suggests to you.
Retention
- Account, profile and learning data — kept while your account is open.
- Coach conversations — kept with the attempt they belong to, so you can resume a mission and review your feedback.
- Work conversations, attached files and deliverables — kept until you delete the conversation, the file or deliverable, or your account. The text read from a Word or Excel file is deleted with the file, and every version of a deliverable with the deliverable. Files removed from a conversation are not included in future requests.
- Work profile and AI kit — kept until you delete them or your account. What you did with each idea is kept while your account is open.
- Work capabilities being tested — what an answer keeps with it (its sources, the code it ran and what that showed, its steps, and the images and files Work made for you) is kept until you delete the conversation or your account, and a long document’s outline and notes until you delete the deliverable, the conversation or your account. A project, with its instructions and files, is kept until you delete it or your account. ZEN’s notes are kept until you delete them or your account, even after the conversation where ZEN kept one is deleted. The steps of an answer that went on in the background are removed after a day (see “Work capabilities being tested”).
- Tools you connect — their tokens are kept until you disconnect the tool or delete your account; the record of each action ZEN took in them, what ZEN asked your permission for and the items shown to you as cards are kept with their conversation, until you delete it or your account.
- What ZEN learns from your work — a conversation’s summary is kept until you delete that conversation, hide it or delete it all in “What ZEN has learnt about you”, or delete your account; the portrait and the suggestions, until you delete it all or your account.
- Usage and AI cost records — kept so we can measure the product over time. When you delete your account they are no longer linked to it, but the usage your browser reported until October 2026 keeps the random identifier it carried (see “Deleting your account”).
- Backups — copies of the database, without the content of the files you upload, taken every day and before updates of the service, each kept for up to 30 days. What you delete stays in them until they expire.
- Contact requests — 12 months from the last time you sent us the form (see “If you ask us to contact you”).
- The newsletter — until you unsubscribe; an address that is never confirmed, 30 days from the last confirmation email (see “Our newsletter”).
Deleting your account
You can delete your account at any time from your profile page. This permanently removes your profile, submissions, coach conversations, scores and certifications, your Work conversations, files and deliverables, your Work projects with their instructions and files, your Work profile, what ZEN remembers about you, the images and files Work made for you, what ZEN learnt from your work and what it suggested to you, your AI kit with what you did with each idea, and the tokens of the tools you connected. It does not always reach those tools: to be sure Rumbo’s access ends there too, disconnect them in Work first, or remove it from each tool.
Some records stay, no longer linked to your account: the usage your browser reported until October 2026, with the random identifier it carried (and, for what it reported before you signed up or while signed out, with all its details); the comments you sent us through the product, and your answers to any pilot survey or interview; the log of actions we took on your account from our administration tools; and, if a company invited you, that invitation with your email address. Other usage records and the AI cost records stay with nothing that identifies you.
Backups keep what you delete for up to 30 more days.
If you ask us to contact you
Two forms on our website need no account: the one that asks for a proposal for your organisation, and the one that, at the end of the four-minute benchmark, asks for the structure of a pilot. From them we keep your email address and what else you fill in: your name, your organisation and, in the benchmark’s form, your role, the size of your finance team and your benchmark score, as well as what you tell us you need and the date you sent it, together with where we stand in answering you and any note we add. We use it only to answer you and to follow up your request.
We process it on the basis of our legitimate interest in answering people who ask us about Rumbo. You can object at any time by writing to soporte@hellorumbo.com.
We read it in our own administration tools. It passes through the servers that run Rumbo (Vercel) and is kept in our database (Supabase); if we write back, our exchange is kept in our email (Google Workspace). It is not sent to any AI provider.
We keep it for 12 months from the last time you sent us one of these forms, and then delete it automatically; backups keep it for up to 30 more days. If you send a form again with the same email address, the new one replaces what you sent before. You can ask us to delete it sooner by writing to soporte@hellorumbo.com. Deleting a Rumbo account does not delete it, because it is kept apart from accounts. If we go on to talk with your organisation about working together, the contact details we keep for that conversation are not deleted with the request: they stay until you ask us to delete them.
Our newsletter
Our newsletter needs no account. When you subscribe we keep your email address, the language you read it in, the area of work you choose if you choose one, the page you subscribed from, the version of the text you agreed to and when, and whether and when you confirmed your address. We use it only to send you the newsletter, about doing the tasks of your job with AI, and the email that asks you to confirm.
We process it on the basis of your consent, which you give when you subscribe and confirm from the email we send you. Until you confirm, we send you nothing else.
It passes through the servers that run Rumbo (Vercel), is kept in our database (Supabase) and is sent through our email provider (Resend). It is not sent to any AI provider. If you reply to one of our emails, your reply is kept in our email (Google Workspace).
Every email carries a link to unsubscribe. When you unsubscribe we delete your address from the list; backups keep it for up to 30 more days. An address that is never confirmed gets nothing else from us and is deleted 30 days after the last confirmation email we sent it. Deleting a Rumbo account does not unsubscribe you, and unsubscribing does not delete your account, because the list is kept apart from accounts.
Your rights
You have the right to access, correct, export and delete your personal data, to ask us to restrict its processing and to object to it. Where we process your data on the basis of your consent, you can withdraw it at any time, without affecting the processing carried out before. Write to soporte@hellorumbo.com. You can also complain to the Spanish Data Protection Agency (Agencia Española de Protección de Datos, www.aepd.es) or to the data protection authority of the country where you live.
Processors
- Supabase — database, authentication and file storage.
- Vercel — application hosting.
- Anthropic — AI coaching and evaluation, the draft of your Work profile and, if you switch it on, the summaries and suggestions of “What ZEN has learnt about you”; Work responses when enabled, with its web search and its environment for running code where you have them.
- OpenAI — Work responses when enabled and selected, with its web search and its environment for running code where you have them; Work’s images where you have them, whichever model answers.
- Resend — transactional email and our newsletter.
- Sentry — error monitoring. Configured to exclude submission content and credentials.
- GitHub — storage of the daily backups of our database.
- Google — our own email (Google Workspace), including what you send to soporte@hellorumbo.com.
- Stripe — payments, only if you buy a paid plan. Nothing is charged while Rumbo is being tested.
If you sign in with your Google account, Google confirms who you are and gives us the account data described under “What we collect”.
Transfers outside the European Economic Area
Most of our providers are based in the United States, and your data is stored and processed there: our database, authentication and file storage (Supabase) and the servers that run Rumbo (Vercel) are in the United States. Each transfer relies on a safeguard recognised by the European Commission: the EU-US Data Privacy Framework, for providers certified under it, or the standard contractual clauses in the provider’s data processing terms. To find out which one applies to each provider, or to get a copy of it, write to soporte@hellorumbo.com.