Privacy notice

What we record when you visit, ask for access or sign in, what sits in your browser, who processes it, and what you can do about it. Short and in plain words, on purpose.

Who we are

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.

What our servers record

Every request is logged with your IP address truncated before it is written (the last octet of an IPv4 address, the last 80 bits of an IPv6 address), the route it hit, the time, and the family of your browser. The truncated address is kept as it is: we do not look up a country or a company from it. Searches, with what was searched for, downloads, exports and matching jobs are recorded the same way, so we can see which features are used and what people look for, and keep the service up.

Every request also carries a session id: a random identifier, created when the page loads and held in the tab's memory, that ties the requests of a single visit together. It is not stored on your device, except in the one case described at the end of this paragraph. It is written to the log whatever you answer to the bar, because it is how one visit is told from another when nothing else identifies you. It is replaced by a new one the moment you answer, so through it the pages you read before your answer cannot be joined to what you do afterwards; the one-day code described under Analytics does count you as the same visitor before and after a yes, for that day only. Two facts about the start of the visit do carry over: the kind of page it began on, such as a factor page but never which factor, and the host name of the site that sent you; if you accept analytics storage, the campaign tags described under Analytics carry over too. If you accepted analytics storage and then sign in, the session id also travels inside the sign-in entry described below, so the visit and the account it opens count as one visit. Otherwise it is gone when you close the tab.

Raw events are kept for 90 days. After that they are exported to our own object storage in the EU and deleted from the database. Records that tie events together (an anonymous id, a link to an account) are kept for 13 months after their last activity, then deleted. Daily totals with no id in them are kept indefinitely.

Cookies and local storage

No identifier for your browser is stored until you accept the bar at the bottom of the page. What is stored before that is only what the site needs to work: your theme, your answer to the bar, your sign-in session, a 15-minute reading pass that lets the pages load factors, one receipt per search so your free searches can be counted, and a few short notes about where you were and what you have already been shown. What is stored, and when:

NameKindWhat it doesLifetime
ocai_themeStrictly necessaryRemembers your light or dark theme choice.1 year
ocai_consentStrictly necessaryRemembers your answer to the bar, so it is not asked again.13 months
portal:search-meterStrictly necessaryCounts the free searches you have used, so your allowance survives a reload. It holds one signed receipt per search: the receipt names the search, not your browser, and carries no search text.2 days
portal:reader-sessionStrictly necessaryLets the pages load factors from our servers. It holds a reading pass that we issue only to this site, which works for 15 minutes and only from the network it was issued to. It is random and names no one, and a new pass replaces it every 15 minutes. None is issued while you are signed in, except to finish loading a page after your sign-in has expired.Works 15 minutes; replaced while you browse, or on your next visit
portal:return-toStrictly necessaryHolds the page you were on when signing in interrupted you, so you land back on it afterwards. Written when you start signing in, and deleted the moment you come back.Until you come back from signing in
portal:oidc-probeStrictly necessaryA one-character test write that checks whether this browser lets us keep you signed in from one visit to the next. It holds nothing about you and is deleted in the same instant it is written.Removed immediately
ocai.reloaded-for-buildStrictly necessaryRemembers which version of the site this tab last reloaded itself for. When we publish a new version while a page is open, the tab reloads once to pick it up, and this note stops it reloading twice. It holds a random version id and nothing about you.Until you close the tab
__Secure-ocai_aid, ocai.aidAnalytics, only after you acceptAn anonymous id that links your visits to each other and, if you sign in, to your account, including after you sign out. Set by the server as a cookie and mirrored in local storage, so a cleared cookie jar does not turn you into a new visitor.13 months
Sign-in sessionStrictly necessaryKeeps you signed in: a cookie on our sign-in server, and the browser entry listed below that lets you come back without signing in again. While you sign up, the cookie on our sign-in server also holds the one word for the kind of site that sent you, if we record it (see Analytics).Until you sign out, or 30 days without a visit

Written by the sign-in library

Signing in itself is handled by oidc-client-ts, a library which writes two entries of its own under names it chooses. We do not write them, but they are on your machine, so they are listed here too. They are the browser side of the sign-in session above, and they appear only once you start signing in.

NameKindWhat it doesLifetime
oidc.user:<authority>:<client id>Strictly necessaryHolds the tokens that keep you signed in, in this site's local storage, so reloading the page, opening a new tab or coming back another day does not sign you out. The parts in angle brackets are our sign-in server and this site's identifier, which is what you would see in the name.Until you sign out, or 30 days without a visit
oidc.<state id>Strictly necessaryHolds one sign-in attempt while it is under way: the secret that proves the reply belongs to your request, the page to bring you back to and, if you accepted analytics storage, this visit's session id, its campaign tags, the kind of page it began on and the host name of the site that sent you. The library deletes it the moment you come back. An attempt you abandon stays behind until you start a sign-in or a sign-out more than 15 minutes later, and one that holds your visit is deleted as soon as you withdraw your consent.Until you come back from signing in; if you abandon it, until a sign-in or sign-out you start more than 15 minutes later

Change your choice

Withdrawing consent stops the analytics id from being set and deletes it from this browser. It does not remove you from our records: we keep one line saying that this browser asked not to be counted, and nothing else. That line is what makes the refusal stick, and it is deleted thirteen months after the last time we saw this browser. What this browser recorded before your refusal stays counted, and keeps its ids until the following night, when a job clears them, along with the truncated address and the browser family, from every event we still hold. Events leave the database after 90 days, into an archive we keep as files: a month archived after your refusal is written without your ids, your truncated address or your browser family, a month archived before it keeps them, and nothing expires those files on its own, so clearing that copy is something we do by hand if you ask us to. From 30 October 2026 this choice covers your browser only: it no longer removes your account's id from what a signed-in account does on our servers, which you object to under Accounts instead. A refusal given while signed in before that date still covers what was recorded until then. The bar comes back so you can answer again, and accepting again resumes the counting from that moment: the refusal is lifted for this browser and, for a refusal given while signed in before 30 October 2026, for your account. It does not bring back what has already been cleared.

Analytics

Analytics is first-party and in-house: our own code, on our own servers, feeding our own dashboards. No advertising, no remarketing, no third-party analytics script, no session replay, and no data sold or shared for anyone else's marketing. We count pages, searches, downloads and the steps of a sign-up so we can tell which parts of the site work.

To count how many people visit each day and which parts of the site they use, our server also works 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 replace every day and delete once its day is over. Nothing is stored on your device for it, and neither the full address nor the browser string is ever stored. The code is written on that day's events in our database and leaves the database with them after 90 days; it is never copied into the archive. It is never linked to an account or to the anonymous id, and once the day's random value is deleted, nobody, including us, can work it out again from an address, so it cannot be traced back to you or joined to another day. We use it only for totals, such as visitors per day or visitors who ran a search, never to look at one visitor. Bots get no code and are not counted as visitors. Answering "No thanks" on the bar, or a browser that sends the Global Privacy Control signal, turns it off; after a no, the codes already written for this browser are cleared that night with its other ids. The basis is our legitimate interest in measuring the audience of our site (Article 6(1)(f) GDPR).

The bar is about your browser. Storing the anonymous id on it, and linking the visits it made before you signed up to your account, happen only if you accept. What a signed-in account does on our servers is different: from 30 October 2026, its searches, the factor pages it opens, its downloads, exports, matching jobs and sign-ins are counted with the account's id whatever you answered, because knowing how the service is used is our legitimate interest (Article 6(1)(f) GDPR). Our servers record it as they answer; no script in your browser reads your sign-in for it. You can object, see Accounts.

We also record where a visit came from, so we can see whether people reach us through a search engine, a link on a social network, an assistant such as ChatGPT, or a newsletter. Two things are kept. The first is the host name of the referring site, for example www.google.com: never the full address, and never what was searched for. It is part of the page view itself, so it is recorded for every visit, including one that declines analytics storage. The second is the utm_source, utm_medium and utm_campaign parameters that a shared or advertised link sometimes carries; those are kept only for visitors who accept analytics storage or sign in, and only once, as the first campaign we saw. Both describe the link, not the person.

From 13 November 2026, when you create an account, we also keep 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 e-mail, another website, or another open-climate.ai site. Your browser works the word out from the site you arrived from, and only the word goes to our sign-in server, on the link that opens the sign-up form. While you sign up, it sits in that server's sign-in cookie (see Cookies and local storage). Once the account exists, the word is kept with it for as long as we keep the account. Our server fetches it once from the sign-in server, when the account is created, and never sends it back to your browser. No page, web address, campaign tag or identifier of the visit is kept with it, and it 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 (Article 6(1)(f) GDPR). You can object by writing to hello@open-climate.ai, and we then delete it. It is not sent if you answered "No thanks" on the bar or your browser sends the Global Privacy Control signal.

Forms

A form on this site (early access, talk to us, a file to try) stores what you type: your name, work e-mail, company, role, message, and the file if you attached one. We use it to answer you and to judge the fit. A copy reaches us by e-mail through Mailgun EU and is mirrored to our CRM in Notion.

Your e-mail address is also hashed, and that hash, never the address, is what links the form to your anonymous analytics id, so we can see which pages a lead read before writing to us.

Accounts

Sign-in runs on Keycloak, which we host ourselves, so your password never leaves our servers. You can also sign in with Google, in which case Google gives us your e-mail address and name and nothing else. An account record holds your e-mail, display name, API keys, and what the account did: which jobs ran, when, and with which result.

From 30 October 2026 we use the account record and what the account does for product statistics: which features are used, how often, on how many days, and by which kinds of organisation and role. We read them as totals and segments in our own admin tool, to decide what to build and to see what does not work, not to decide anything about you. The basis is our legitimate interest (Article 6(1)(f) GDPR).

We also add business contact details to an account: your company, its type, your job title, seniority and 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 your work e-mail domain, with the help of an AI assistant. We use them for the same statistics and to manage our relationship with your organisation, never for advertising, and they are kept as long as the account record. The basis is again our legitimate interest.

You can object to both at any time by writing to hello@open-climate.ai and, from 30 October 2026, with the switch on this page when you are signed in. 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. What was recorded before 30 October 2026 keeps the treatment it was recorded under.

Product statistics for your account

Sign in to see and change this setting for your account. You can also object at any time by writing to hello@open-climate.ai.

Who else sees it

We use a small number of providers that process data only on our instructions:

  • OVHcloud (France): hosting, database, object storage, backups.
  • Sentry (EU region): error and performance monitoring, configured not to receive personal data.
  • Mailgun EU: transactional and support e-mail.
  • Google Workspace: our own e-mail and documents.
  • Odoo: customer relationship management.
  • Notion: customer relationship management, being retired in favour of Odoo.
  • Anthropic (United States): the AI assistant we use to look up the business contact details of accounts.

Your rights

You can ask for a copy of your personal data, have it corrected or deleted, restrict what we do with it, object to processing based on our legitimate interest, and withdraw consent at any time. We answer within one month and do not ask you to justify a request.

Write to hello@open-climate.ai.

If you think we have got it wrong, you can complain to the Belgian Data Protection Authority, Rue de la Presse 35, 1000 Brussels, www.dataprotectionauthority.be, or to the authority where you live or work.