Polityka Prywatności
Ostatnia aktualizacja: listopad 2026
Ta polityka wyjaśnia, jakie dane zbieramy, dlaczego, i jakie masz prawa, zgodnie z Rozporządzeniem (UE) 2016/679 (RODO).
Pełny tekst tej strony jest dostępny tylko po angielsku i to wersja angielska jest wiążąca.
1. Data controller
The controller of your data is LoadFocus LTD, trading as Changeloop, a company registered in England and Wales under company number 09554514, with its registered office at 4 Grayscroft Road, London SW16 5UP, United Kingdom. You can contact us at contact@changeloop.dev.
2. What data we collect
Account data (name, email address, and email via GitHub OAuth where you use it to sign in); source integration data (for a GitHub App installation, the installation ID and repository metadata; for a GitLab or Bitbucket connection, the project path, the address of your self-managed GitLab server where you use one, the branch you name and the webhook secret we issue; for every source, the titles and bodies of merged pull requests or merge requests and, if you choose push mode on Bitbucket, the messages of pushed commits, which we use to draft changelog entries and delete once drafting finishes, except that we keep the title of a change we skip or that arrives over your plan's limit, so you can review it; we also use the title and body of a GitHub issue you label for the roadmap to draft its roadmap item, and delete them once drafting finishes, whether or not the draft succeeds; we do not store your source code); billing data (invoices and subscription status from Stripe, for paid Team plan accounts); technical usage data (logs, cookies, including the session cookie that keeps you logged in). If you use our widget or hosted changelog page, we also process data about the people who read your changelog, on your behalf: the message and email address a visitor sends as feedback, the email address of anyone who subscribes to your changelog by email, and emoji reactions and votes on a public feedback board, both recorded against a random browser identifier, never a name or email address (see section 6 below).
3. Purposes of processing
We use the data to draft and publish changelog entries from your merged pull requests, merge requests or pushed commits, manage your account and subscription, provide support and improve the service. Where a customer has connected GitHub and enabled the feedback widget, we also use it to record their visitors' submissions and create the corresponding issue in that customer's GitHub repository. Where a customer has switched these features on, we use it to email their changelog's updates to the readers who subscribed, to tell a person who sent feedback when what they asked for has shipped, to count reactions to their changelog entries, and to count votes on the requests they publish on their feedback board.
4. Legal basis
Processing is based on the performance of the contract (providing the service), on your consent where applicable, and on the legitimate interest of securing and improving the platform. Feedback submitted through the widget is processed under our contract with the customer who enabled it, and on the basis that the person submitting it chose to do so voluntarily. Changelog email subscriptions are processed under our contract with the customer whose changelog it is, and on the basis of the subscriber's consent, which they give by confirming their address and can withdraw at any time with the unsubscribe link in every email. The email telling someone that their request has shipped is processed under our contract with the customer who switched it on, and on the basis that the person was told when they sent the request that they would get it. Emoji reactions are processed under our contract with the customer who switched them on, and on the basis that the reader chose to react. Votes on a public feedback board are processed under our contract with the customer who switched the board on, and on the basis that the reader chose to vote.
5. GitHub access
To provide the service we access your repository metadata and the titles and bodies of merged pull requests through your GitHub App installation, using short-lived installation tokens rather than long-lived personal access tokens. The GitHub App has no permission to read your repository's files. GitLab and Bitbucket send us the same kind of data through the webhook you configure, along with commit messages if you choose push mode on Bitbucket, and we hold no access token for them. If you use Connect with Bitbucket, Bitbucket issues us a token that can read the repository and its pull requests and manage its webhooks (Bitbucket requires the pull request permission for a webhook that reports merged pull requests); we use it within a single request to read the repository's main branch and add the webhook, and we do not store it. We do not store your source code.
6. Data about our customers' visitors and readers
Our customers can put Changeloop on their own websites with our widget, and share their changelog on a page we host for them. When their visitors and readers use these, we process the data below on the customer's behalf. Email subscriptions, emails about shipped requests, emoji reactions and the public feedback board are off until the customer switches them on.
Feedback. A visitor can send a short message and an email address through the widget. We store both, together with a randomly generated submission reference and, once one exists, the GitHub issue it results in. Before the message is used to classify the submission or included in that issue, we automatically remove the submitter's email address from it, so the resulting GitHub issue and our AI classification provider never see it; only the message text does, with the email address stripped out. The person who submitted the feedback can check its status using the private link they were given at submission time.
An email when a request ships. If the customer has switched this on, the widget tells the person sending feedback, under the email field, that they will get an email when their request ships. We email only people who were shown that notice when they sent their request. When the customer publishes a changelog update that delivers the request, we send one email to the address given with it: one for each request, and no more than one per person for each update, however many of their requests it delivers. The email names the customer's changelog and the update, links to it, and does not repeat the message the person sent. We record when it was sent with the feedback. Every email has a link to stop these emails from that customer. We remember that choice as a hash of the address, combined with the customer's team, rather than the address itself.
Changelog email subscriptions. If the customer has switched this on, a reader can ask, in the widget or on the hosted changelog page, to get the customer's updates by email. We store the email address and send an email asking the person to confirm it, at most one a day. Its link works for 48 hours, and nothing else is sent until they confirm. If the address is not confirmed within 7 days, we delete it. Once it is confirmed, we send at most one email for each update the customer publishes, and keep the address, when it was confirmed and the name of any delivery error. Every email has an unsubscribe link, and mail apps that offer one-click unsubscribe can use it. When someone unsubscribes, we delete their address and keep only a hash of it combined with the customer's team, the dates, and the code from their unsubscribe link. The customer can see how many people have subscribed and download a list of the confirmed addresses.
Emoji reactions. If the customer has switched this on, readers can react to a changelog update with any of five emoji, in the widget or on the hosted changelog page. The first time someone reacts, their browser creates a random identifier and keeps it in its local storage. We receive that identifier with each reaction and store only a hash of it combined with the customer's team, so the same browser leaves unrelated hashes on different customers' changelogs. It is never linked to an email address or an account. We use it only so that each browser counts once for each reaction, and the customer sees counts only, never who reacted. Clearing the browser's storage deletes the identifier, and the browser gets a new one the next time it reacts.
Votes on a public feedback board. If the customer has switched this on, the widget and the hosted changelog page show a board of requests the customer chose to publish, and readers can vote for them. The customer writes each request's title and description. Starting a request from a visitor's feedback does not publish that feedback: the board never shows the visitor's message or email address. A vote uses the same random identifier as reactions, which makes it pseudonymous: it does not tell us who the reader is, but it is still personal data. We store only a hash of the identifier combined with the customer's team, made differently from the reaction hash, so that a stored vote hash and a stored reaction hash cannot be matched to each other, together with the request and when the vote was recorded. It is never linked to an email address or an account. We use it only so that each browser counts once for each request, and the customer sees vote counts only, never who voted. Taking a vote back, which the reader can do while the request is on the board, deletes it. The browser also keeps a list of the requests it voted on (see section 12).
We do not store IP addresses with subscriptions, reactions or votes; to limit abuse, we count requests per IP address in short-lived counters only. We send these emails through Amazon Simple Email Service (see section 8). All of this data is included in the copy of their data a customer can download from Settings, without the hashes or the codes in email links, and with reactions and votes as counts only. It is permanently deleted when the customer deletes their account (see section 10).
For this data we act as a processor on behalf of the customer whose widget or changelog was used; that customer is the controller, and a request about this data should usually go to them, though we are happy to help too. The same applies to other personal data a customer has us process for them, such as names that appear in their pull requests: our Data Processing Agreement, which forms part of our Terms, sets out how we handle it.
7. Storage and security
Data is stored on secure servers and protected by reasonable technical and organisational measures. Access is limited to the purposes described above.
8. Data sharing
We do not sell your data. We share personal data only with the service providers below, who process it for us under contracts that require them to keep it secure and to use it only to provide their service to us, or where the law requires us to.
- Amazon Web Services: hosting, our databases and their backups, our content delivery network (Amazon CloudFront), and sending email (Amazon Simple Email Service). Our servers are in the United States, in the AWS us-east-1 region (Northern Virginia).
- OpenAI: the AI that drafts changelog entries from the titles and descriptions of your merged pull requests or merge requests (and from commit messages in Bitbucket push mode), along with their alternative versions, translations and roundups, roadmap items from the GitHub issues you label, and tag suggestions. It also classifies messages sent through the feedback widget, after we have removed the submitter's email address (see section 6). OpenAI processes this data in the United States.
- Stripe: payments for the Team plan. Card details go from your browser straight to Stripe, through Stripe's own payment form, and never reach our servers. We give Stripe your email address, name and billing details to set up your subscription, and Stripe tells us whether each payment succeeded. Stripe also uses some of this data as a controller in its own right, for example to prevent fraud, under its own privacy policy, and may process it in the United States.
- Google: reCAPTCHA on our sign-up page, which checks that a sign-up comes from a person rather than a bot and, to do that, collects information about your browser and how you use the page; sign-in with Google, if you choose it, in which case Google confirms your email address to us; and Google Workspace, which hosts our contact@changeloop.dev mailbox, so the emails you send us are stored by Google. Google may process this data in the United States.
- Our page speed monitoring: our pages, our blog and the Changeloop app load a small script from speed.loadfocus.com, the page speed service run by LoadFocus LTD, the same company that runs Changeloop, on Amazon Web Services in the United States. It sets no cookies. It reports the page's address without any query string and timings such as how quickly the page showed its main content, and from the request we work out your country, browser and whether you are on a mobile or desktop device. These are kept only as totals. Your IP address is used only for that and to limit abuse, in short-lived counters, and is never stored with the measurements.
Services you connect are different. GitHub, GitLab and Bitbucket send us your pull requests, merge requests, labelled issues or commits, and GitHub receives the issues we create from widget feedback, because you connected them; Slack receives the entries you choose to post to your channel; an AI assistant you connect to our MCP server reads your changelog with the key you give it; and if you sign in with GitHub, GitHub confirms your identity to us. These services receive or send data only because you connect them, and they act under their own terms and privacy policies, not as our processors.
If you are a customer and use Changeloop to process personal data on your behalf, our Data Processing Agreement applies, and the providers above that handle that data are its sub-processors. We will update this list at least 30 days before a new provider starts processing such data, so that you can object.
9. International transfers
We are based in the United Kingdom, and the data we hold is stored in the United States, on Amazon Web Services' servers in Northern Virginia. Some of the providers in section 8 also process data in the United States. When personal data is transferred out of the United Kingdom or the European Economic Area, we rely on the EU-US Data Privacy Framework and its UK Extension where the provider is certified under them, and otherwise on the European Commission's Standard Contractual Clauses together with the UK Addendum issued by the Information Commissioner. You can ask us for a copy of the safeguards that apply at contact@changeloop.dev.
10. Retention period
We keep your account data for as long as you have an account, and other data only for as long as we need it for the purposes in section 3, or for longer where the law requires it.
When you delete your account, we:
- cancel any subscription straight away and remove your saved cards from Stripe;
- delete your sign-in sessions and tokens, password reset links, API key and usage limits, the reason you gave if you moved to the Free plan, and our records of failed-payment reminders and of Stripe's notifications about you;
- anonymise your account record: we remove your name, email address, password, the GitHub or Google account you signed in with, and your profile and contact details. What is left is a random account ID, the date you deleted the account and, if you paid us, the Stripe customer reference that links your payment records. Any other session you had open stops working;
- through an automated process that runs every hour, permanently delete the data of the teams you own, as described in section 12 of our Data Processing Agreement, and remove our GitHub App installation from your account.
If we issued you an invoice, we keep the invoices and the records of your payments, with the billing details they rely on: your name, email address, postal address, company and VAT number, and the evidence of your country that VAT rules ask for (your card's country, brand and last four digits, and the IP address you paid from). We keep these, and the customer record we created at Stripe, which holds the same payment history, for ten years from the end of the financial year they belong to, which covers both UK tax rules and the EU's VAT rules. If you never paid us, your billing details and the customer record we created at Stripe are deleted with the rest.
Copies in our backups are not changed one by one. They expire on a rolling schedule, the last of them no later than 12 months after it was taken, and until then they stay encrypted and are used only to restore the service.
11. Your rights (GDPR)
You have the right of access, rectification, erasure, restriction, portability, and objection, as well as the right to withdraw your consent. To exercise them, write to contact@changeloop.dev. You also have the right to lodge a complaint with your local data protection supervisory authority. If you sent feedback, subscribed to updates, reacted or voted through one of our customers' widgets or changelog pages rather than holding an account with us, that customer is usually the right party to contact, since they are the controller for that data; you are welcome to write to us as well and we will help.
12. Cookies
We use only cookies that are strictly necessary for the site to work: expresscook.sid, our session cookie, which expires after 7 days, and, once you sign in, session_token, which keeps you signed in for up to 14 days. We do not use analytics or advertising cookies, and the page speed script described in section 8 sets none. The Changeloop app keeps a few display preferences, such as whether the sidebar is collapsed, in your browser's local storage. The widget on our customers' sites keeps a few things in the visitor's local storage: for each changelog, when the visitor last saw it, so the widget can count what is new and show its release pop-up only once; the references of any feedback they sent, so it can show the status of their requests; if they react to an update, the reactions they chose; and, if they vote on a feedback board, the requests they voted for, under changelogapp-widget:votes: followed by the changelog's id. Reactions and votes share the random identifier described in section 6, kept under changelogapp-widget:reader-id, which is one per browser on each website. Our hosted changelog page keeps the same reaction and vote details in local storage under its own address. The widget sets no cookies. Google sets its own cookies when reCAPTCHA runs on our sign-up page or when you sign in with Google, and Stripe may set cookies to prevent fraud on pages where its payment script runs. When a customer turns on images in their changelog, readers' browsers and mail clients load those images directly from the hosts the customer lists, for example GitHub, without sending a referrer. That host receives the request like any other image request, under its own privacy policy; we receive nothing from it.
13. Changes and contact
We may update this policy; changes will be published here. For any question regarding your data: contact@changeloop.dev.