1. What this policy is for
This policy explains what we do with personal data when you visit open-climate.ai, ask us for access, or use the Service with an account. It is written for the people whose data it is: visitors, prospects, and the individuals who use the Service on behalf of a customer.
It does not cover the content our customers send us for processing. When a customer uploads a procurement file or sends us activity lines, we act on that customer's instructions, not our own, and the Data Processing Agreement governs it. Article 4 below explains where the line falls.
2. Who is responsible
2.1 The controller is Songbird AI BV, Rue Emile Wittmann 50, 1030 Schaerbeek, Belgium, enterprise and VAT number BE 1033.450.866. open-climate.ai is its trading name.
2.2 For anything in this policy, write to hello@open-climate.ai.
2.3 We have not appointed a Data Protection Officer. Our core activity is not the large scale monitoring of individuals, nor the large scale processing of special categories of data, so Article 37 GDPR does not require one. If that changes, we will appoint one and say so here.
3. What we collect, why, and on what basis
3.1 When you visit the site
Our servers log the request: IP address, the page or endpoint, the time, the user agent, and the response. We need these to serve the page, to keep the Service up, and to detect abuse, which is our legitimate interest under Article 6(1)(f) GDPR. Logs are kept for 12 months.
From 6 November 2026, our servers also work out a one-day code for each visit, so we can count how many people visit each day and which parts of the site they use. The basis is our legitimate interest in measuring the audience of our site, Article 6(1)(f) GDPR. Nothing is stored on your device for it, it is used only for totals, and a "No thanks" on the banner or a browser that sends the Global Privacy Control signal turns it off. Article 6.2 says how it works and how long it is kept.
When you are signed in, our servers also record what your account does on them: the searches it runs, the factor pages it opens, the downloads, exports and matching jobs it starts, and when it signs in. We use that usage data for the product statistics described in article 3.3. The basis is our legitimate interest, Article 6(1)(f) GDPR, and you can object to it at any time.
From 13 November 2026, when you create an account, we record on it one word for the kind of site that sent you: none (you typed our address or used a bookmark), a search engine, a social network, an AI assistant, an email, another website, or another open-climate.ai site. Your browser works that word out from the site you arrived from, and only the word is passed to our sign-in service, on the link that opens the sign-up form. We record nothing else about that visit: no page, no web address, no campaign tag and no identifier, and the word is not linked to anything else. We use it only to count new accounts by the kind of site that sent them. The basis is our legitimate interest in knowing which channels bring us accounts, Article 6(1)(f) GDPR, and you can object to it at any time. If you answered "No thanks" on the banner, or your browser sends the Global Privacy Control signal, the word is not passed and nothing is recorded. Article 6.2 says what sits on your device while you sign up, and how long the word is kept.
Two things need your consent instead, Article 6(1)(a) GDPR: storing an analytics identifier on your device, and using it to link the visits you made before you had an account to that account. You give or refuse that consent through the banner described in article 6.2, and you can withdraw it at any time.
3.2 When you ask us for access
The request form collects your name, your work email, your company, your role, optionally the tool you use today, your message, and optionally a file you attach so we can show you what the Service does with it. We use it to answer you, to evaluate the fit, and to keep in touch about the product. The basis is our legitimate interest in responding to a business enquiry and in business to business marketing, Article 6(1)(f) GDPR, and taking steps at your request before entering a contract, Article 6(1)(b) GDPR.
The form also carries a hidden field that only automated submitters fill in. If it is filled in we discard the submission. That is spam protection, not profiling.
We keep prospect records for 24 months after our last meaningful contact, and delete them after that unless you have become a customer. You can ask us to stop contacting you at any time, and we will act on it without asking for a reason.
3.3 When you use the Service with an account
We process your identity (email address, display name, and either a password that we store only as a hash or, where you sign in through your own identity provider, the identifier that provider gives us), the API keys and MCP connections issued to you, and a record of what your account did: which jobs ran, when, from which address, and with which result. We need this to give you the Service and to bill for it, Article 6(1)(b) GDPR, to keep it secure and to investigate misuse, Article 6(1)(f) GDPR, and to meet our accounting obligations, Article 6(1)(c) GDPR.
We also use the account record, and the usage data in article 3.1, for two purposes of our own. Both rely on our legitimate interest, Article 6(1)(f) GDPR, and neither is a condition of using the Service.
Product statistics. We count how accounts use the Service: which features, how often, on how many different days, and how that differs between kinds of organisation and kinds of role. We need it to decide what to build, to see what does not work, and to report on the business. We read it as totals and segments in our own administration tool, on our own servers. It is recorded by our servers as they answer your account's requests; no script in your browser reads your sign-in for it.
Business contact details. We add to an account the business details that describe it: your company, the type of company it is, your job title, your seniority, your function, and a link to your public professional profile. We find them in public professional sources, such as company websites and professional networking profiles, starting from your name and the domain of your work email, with the help of the AI assistant listed in article 5. We use them for the product statistics above and to manage our relationship with your organisation. We do not use them for advertising, we do not sell them, and no decision about you is taken from them by automated means. For the details we did not receive from you, this paragraph is the notice Article 14 GDPR requires.
You can object to either purpose at any time under Article 21 GDPR, with the switch in the signed-in part of our privacy page or by writing to hello@open-climate.ai. From the moment you object, what your account does is no longer attributed to it, and we delete the business contact details we added and stop adding new ones. We do not claim compelling grounds that would override an objection to either purpose.
These two purposes apply from 30 October 2026. Usage recorded before that date keeps the treatment it had under version 1.2 of this policy, and is not attributed to an account after the fact.
Account and access records, the business contact details and the kind of site that sent you (article 3.1) included, are kept for as long as the account is open, and for 12 months after it closes. Usage data is kept as article 6.2 states. Accounting records are kept for 7 years, as Belgian law requires.
3.4 When something breaks
The Service reports errors and performance traces to our monitoring tool so we can fix them. It is configured not to attach personal data to those reports. A stack trace can still incidentally contain an identifier, so we treat the monitoring data as personal data and apply the same rules to it. The basis is our legitimate interest in a Service that works, Article 6(1)(f) GDPR. Retention is 90 days.
3.5 When you write to us
Email you send us, and our replies, are kept for 24 months so we have a record of what was agreed and what was asked. The basis is our legitimate interest in managing the relationship, Article 6(1)(f) GDPR. Our email runs on Google Workspace, listed in article 5.
3.6 When we contact you first
We also approach people we think the Service is useful to. If we contact you without your having contacted us first, this section is the notice Article 14 GDPR requires, and you are reading it at or before our first message.
Where it comes from. Professional profiles you have made public, your employer's website, public company registers, and, if we have met, what you told us.
What we hold. Your name, job title, employer, the public address of your profile, the location it states, your work email and phone number when you have made them available to us, our notes on the conversation, and the messages exchanged between us.
Why, and on what basis. To assess whether your organisation has a use for the Service and to contact you about it in your professional capacity. The basis is our legitimate interest in business to business marketing, Article 6(1)(f) GDPR. We have weighed that against your interests: the data is professional rather than private, it concerns you in your working role and not your private life, we do not build a profile or score you, we do not take any decision about you by automated means, and one word from you ends it.
How long. 24 months from our last meaningful contact, and immediately if you ask us to stop.
What you can do. Object at any time under Article 21 GDPR, by replying to any message or writing to hello@open-climate.ai. We will stop and delete the record without asking you for a reason. Article 9 sets out your other rights.
4. Customer content is different
When a customer uses the Service to map their own data, the files and lines they send us may contain personal data: a supplier's contact name, an employee named on a travel line, an approver in a procurement system. We process that content only to run the Service, only on the customer's instructions, and we do not decide what happens to it. In GDPR terms the customer is the controller and we are the processor.
Two consequences follow. If you are an individual whose data appears in a customer's file and you want it accessed, corrected or deleted, ask that customer, and we will help them answer you. And the rules that govern that content, including our commitments on model training, on where it is processed and on how long it is kept, are in the Data Processing Agreement and in articles 7.3 to 7.6 of the Terms of Service, not in this policy.
5. Who else sees it
We do not sell personal data, and we do not share it for anyone else's marketing. We use a small number of service providers, and they process data only on our instructions.
| Provider | What it does | Where the data sits |
|---|---|---|
| OVHcloud | Hosting, database, object storage, backups | France |
| Sentry (EU region) | Error and performance monitoring | Germany |
| Mailgun EU | Transactional and support email | European Union |
| Google Ireland (Workspace) | Our own email and documents | European Union, with transfers to the United States |
| Odoo SA | Customer relationship management: prospect and customer records | European Union |
| Notion Labs | Prospect and contact records, and our internal working notes | United States |
| Anthropic | The AI assistant we use to look up the business contact details in article 3.3 | United States |
Neither of the two customer relationship tools ever receives customer content. The files and lines a customer uploads to the Service stay on the infrastructure in the first row of the table; article 4 and the Data Processing Agreement govern them, and no business tool listed here has access to them. What those two rows hold is contact and prospect records: names, job titles, employers, work email addresses and phone numbers, and our own notes on the conversation.
Our authentication runs on software we host ourselves, on the infrastructure above, so signing in does not hand your credentials to a third party. So does our site analytics: the usage data in article 6.2 goes to our own servers on the infrastructure above and to no analytics company, so there is no row for it in this table.
The model provider that powers the Service processes customer content, not the data in this policy. It is listed in Annex III of the Data Processing Agreement, and it processes in the European Union. The AI assistant in the last row of the table is a different provider with a narrower job: it receives the name, the work email domain and the company of an account when we look up that account's business contact details, and never any customer content.
We also disclose personal data where the law requires it, and to our professional advisers under a duty of confidentiality. If we are ever part of a merger or a sale of the business, data moves with it, and we will tell you before that happens.
6. Cookies and analytics
6.1 Strictly necessary cookies. Signing in sets a session cookie, forms carry a token that prevents cross-site request forgery, and long running requests carry a cookie that keeps you on the same server for the life of the connection. While you sign in, our sign-in service also keeps the progress of that sign-in in a cookie of its own, and when you create an account that cookie holds the one word described in article 6.2. These are needed for the Service to work, we do not ask consent for them, and Article 129 of the Belgian Act of 13 June 2005 does not require it.
6.2 Analytics. We count visits and see which pages are read with our own software, on our own servers. There is no third-party analytics product, no advertising network and no tag manager: the page sends an event to our API, and it stays in our database on the infrastructure in article 5.
The banner is about your device. Storing an identifier on it, recognising the
same browser from one day to the next, and linking the visits it made before
you had an account to that account are the parts that are not strictly
necessary, so they are the parts we ask about. Until you answer the banner, and
if you decline it, nothing persistent is stored for your visit. Before 6
November 2026 such a visit is counted without any identifier. From that date, a
visit you have not yet answered the banner for also gets the one-day code
described below, and a visit for which you declined is still counted without
any identifier. If you accept, we store a random identifier that says nothing about
you: a __Secure-ocai_aid cookie our server sets, and a copy of the same value
in your browser's local storage in case the cookie is cleared. It is a number,
it is not linked to your name unless you later give us one by signing in or
writing to us, and we use it only to tell a returning visit from a new one and,
once you have given us your name that way, to link this browser's earlier
visits to it.
What a signed-in account does on our servers is not decided by the banner. From 30 October 2026 it is recorded with the account's identifier, whatever you answered, for the product statistics in article 3.3 and under our legitimate interest as articles 3.1 and 3.3 explain. You object to it with the switch in the signed-in part of our privacy page or by email, not through the banner. Declining the banner while signed in still deletes the identifier from this browser and stops the linking of its earlier visits; it no longer removes the account's identifier from what the account does. A refusal given while signed in before 30 October 2026 keeps covering what was recorded before that date.
The kind of site that sent you. From 13 November 2026, one fact about the
visit in which you create an account reaches that account without your consent:
the one word for the kind of site that sent you, described in article 3.1. It
is the only exception to the rule above that the visits made before you had an
account are linked to it only if you accept the banner. Your browser works the
word out from the site you arrived from and adds it to the link that opens our
sign-in service. That service copies the details of the sign-in request into
its own strictly necessary sign-in cookie, called KC_RESTART, on
auth.open-climate.ai, so for the length of the sign-in the word sits on your
device inside that cookie. Once the account exists, the word is kept on it, in
our sign-in service and in our account record. Our server fetches it once from
the sign-in service, when the account is created, and it is never sent back to
your browser. It is kept for as long as the account record, as article 3.3
states. The word is not
sent if you answered "No thanks" on the banner or your browser sends the Global
Privacy Control signal; if you have not answered the banner, it is sent. To
object, write to hello@open-climate.ai and we delete it from your account.
Accounts created before 13 November 2026 have no such word.
Daily visitor count. From 6 November 2026, to count how many people visit each day and which parts of the site they use, our servers work out a one-day code for each visit. It is computed from your IP address and the string your browser identifies itself with (its user agent), mixed with a random value that we create for each day, counted in Coordinated Universal Time, and delete once that day is over. Nothing is stored on your device for it, nothing is read from your device beyond what your browser sends with every request, and neither the full address nor the user agent is stored for it. The code is written on that day's events in our database and is deleted with them after 90 days; it is never copied into the archive described below. It is never linked to an account, to the analytics identifier above, or to the code of another day. Once the day's random value is deleted, nobody, including us, can work a code out again from an address, so a code cannot be traced back to you or joined to another day. We use the codes only for totals, such as visitors per day or visitors who ran a search, and never to look at one visitor. Traffic we recognise as automated gets no code and is not counted as a visitor. The basis is our legitimate interest in measuring the audience of our site, Article 6(1)(f) GDPR. A "No thanks" on the banner, or a browser that sends the Global Privacy Control signal, turns it off: no code is computed for your visits from then on, and after a "No thanks" the codes already written for this browser are cleared that night with its other identifiers.
Raw events are kept in our database for 90 days, and then in an archive of files on our own storage in the European Union, without the one-day codes. Records that tie events together, an analytics identifier or its link to an account, are kept for 13 months after their last activity. Daily totals with no identifier in them, the daily visitor counts included, are kept without a time limit.
We remember your answer in a first-party cookie called ocai_consent for
thirteen months. That cookie needs no consent of its own, because remembering
that you declined is what honouring it means. You can change your mind at any
time through the cookie settings on our privacy page, which brings the banner
back.
6.3 We do not use what this article describes to build a profile of you for advertising or to target you, and we do not currently use advertising cookies. If we start, they will sit behind the same consent gate as article 6.2, we will list them here before we set them, and declining costs you nothing. The one-day code in article 6.2 counts visitors and nothing else, and the kind of site that sent you is used only to count new accounts by channel. The product statistics and business contact details in article 3.3 describe how accounts use the Service and who they work for; we read them as totals and segments and use them to manage our customer relationships, and article 10 applies to them.
7. Where your data is
The Service itself runs entirely in the European Union. Our servers, our database, our object storage, our backups, our monitoring and the model processing behind the Service are in EU data centres, and our staff access them from within the EU or the EEA.
Site analytics is part of that: it runs on our own servers, so nothing about your visit leaves the European Union for it.
Three of the business tools in article 5 are the exception: our email and documents on Google Workspace, the contact records and working notes on Notion, and the AI assistant on Anthropic that looks up the business contact details of accounts. None of them holds customer content. Those transfers rely on the EU to US Data Privacy Framework where the recipient is certified under it, together with the European Commission's standard contractual clauses.
Where we ever need to transfer personal data outside the EEA for another reason, we will do it under an adequacy decision or under the European Commission's standard contractual clauses, and we will assess whether the law of the destination country actually lets those safeguards work before we rely on them.
8. How we protect it
Traffic is encrypted in transit with TLS. Data at rest, including backups, is encrypted. Access is role based and granted on need, credentials and secrets are held encrypted, file downloads are re-checked against an access token on every request rather than exposed as public links, and we keep an access log we can reconstruct an incident from. Annex II of the Data Processing Agreement sets out the measures in more detail.
No system is perfectly secure. If a breach affects your personal data and is likely to put you at risk, we will tell the Belgian Data Protection Authority within 72 hours and tell you without undue delay.
9. Your rights
Under the GDPR you can ask us to give you a copy of your personal data, correct it, delete it, restrict what we do with it, or hand it to another provider in a portable format. Where we rely on legitimate interest, you can object, and we will stop unless we have compelling grounds that override your interests. For the product statistics and business contact details in article 3.3 we claim no such grounds: an objection, with the switch in the signed-in part of our privacy page or by email, ends both for your account from the moment it is set. For the daily visitor count in article 6.2 we claim none either: a "No thanks" on the banner, or the Global Privacy Control signal, is your objection, and from then on no code is computed for your visits. Once a day is over we can no longer tell which of that day's codes were computed for your visits, so we cannot single them out to give you a copy or to delete them (Article 11 GDPR). For the kind of site that sent you, in articles 3.1 and 6.2, we claim none either: write to us and we delete it from your account. Where we rely on consent, you can withdraw it at any time, which does not affect what we did lawfully before.
Write to hello@open-climate.ai. We answer within one month, and we will tell you if we need longer because a request is complex. We do not charge for this, and we do not make you justify a request to stop marketing.
If you think we have got it wrong, you can complain to the Belgian Data Protection Authority, Rue de la Presse 35, 1000 Brussels, contact@apd-gba.be, www.dataprotectionauthority.be, or to the authority where you live or work.
10. Automated decisions
We do not take decisions about individuals by automated means that have legal effects on them or affect them in a similarly significant way, so Article 22 GDPR does not apply to what we do. The Service classifies and matches business activity data, not people. The product statistics and business contact details in article 3.3 group accounts by how they use the Service and by the kind of organisation and role they belong to; we use those groups to understand the product and to manage our customer relationships, and no decision about you is taken from them by a machine.
11. Children
The Service is for organisations, not individuals acting privately, and it is not directed at children. We do not knowingly collect their data.
12. Changes
We may update this policy. If a change matters to you, we will say so by email or on the site before it takes effect, and the version and date at the top will change. We keep previous versions and will send you any of them on request to hello@open-climate.ai.
Version 1.3, in force from 30 October 2026, changes the basis on which we count what a signed-in account does on our servers from consent to our legitimate interest (articles 3.1, 3.3 and 6.2), adds the business contact details we add to an account from public professional sources (article 3.3) and the AI assistant that looks them up (articles 5 and 7), and says how to object to both. The banner now governs only what is stored on your device and the linking of visits made before you had an account. Usage recorded before 30 October 2026 keeps the treatment it had under version 1.2.
Version 1.4, in force from 6 November 2026, adds the daily visitor count: a one-day code our servers work out from your IP address and user agent, under a random value deleted at the end of each day, to count visitors per day under our legitimate interest. Nothing is stored on your device for it, it is never linked to an account, to the analytics identifier or to another day, and it is used only for totals. A "No thanks" on the banner or the Global Privacy Control signal turns it off (articles 3.1, 6.2, 6.3 and 9). Visits before 6 November 2026 have no such code.
Version 1.5, in force from 13 November 2026, records on a new account one word for the kind of site that sent you, such as a search engine or a social network, under our legitimate interest and without your consent (articles 3.1, 6.1, 6.2, 6.3 and 9). Nothing else about that visit is kept, and the word is not linked to anything else. While you sign up, the word sits in our sign-in service's strictly necessary sign-in cookie. A "No thanks" on the banner or the Global Privacy Control signal stops it being sent, and you can object by email. Accounts created before 13 November 2026 have no such word.
13. Language
This policy is drafted in English. Where we publish a translation of it, or of the privacy notice on our site, and the versions differ, the English version prevails.
14. Contact
Songbird AI BV Rue Emile Wittmann 50, 1030 Schaerbeek, Belgium Enterprise and VAT number BE 1033.450.866 Privacy, and everything else: hello@open-climate.ai