Skip to content

How Arkiviva processes data on customer instructions

Last updated: 16 November 2026.

Our role

When you store content with Arkiviva, such as repositories, issues, wikis, packages, and Pages sites, you decide what happens to any personal data in it, whether the content is public or private. We process that data on your instructions, as your processor, under the data processing agreement between you and Arkiviva, apart from a few uses of our own that our Privacy Notice describes, such as protecting the service. If you are yourself a processor for another organisation, we are your subprocessor. This page explains that processing in plain terms; the data processing agreement sets the terms.

Our own uses of personal data, such as running your account and keeping the service secure, are described in our Privacy Notice.

What we do on your instructions

Service Personal data that may be involved What we do
Organizations Memberships, roles, permissions, and invitations that you set up Apply your access rules and send the invitations you ask for.
Repositories and collaboration Content that contains personal data; names and email addresses in Git history; issues, pull requests, reviews, comments, wikis, attachments, and activity Store, display, search, and share content according to your settings. People can find public repositories within Arkiviva.
Packages, releases, large files, imports, and integrations Files and their metadata; imported authorship and collaboration records; connection details you give us Store or transfer what you select, connect to the services you set up, and send the webhooks you configure.
Pages Published files and their source; your domain name and the information needed to validate it; requests from visitors to your site, including their IP addresses, while we serve them Build and serve your site, validate your domain, and obtain its certificates. Certificates are recorded in public certificate logs, so your domain name becomes public.
Actions Workflow definitions, job inputs, logs, outputs, artifacts, and secrets you give a job Run the workflows you request. Runners you bring yourself run on infrastructure you choose.
Running the service Your data, as needed for diagnostics, backups, and restored copies Maintain, update, troubleshoot, and back up the service, and restore backups to separate systems to check recovery and to test updates.

Public content can be copied, indexed, forked, or cloned by others, and those copies stay outside our control.

How long we keep your data

Data How long we keep it
Your content and the personal data in it While you use the service. After you delete it, or your account or the service ends, we delete it from our live systems within 90 days, or within 90 days after any time you have to retrieve it.
Backups and restored copies Deleted within 13 months after the data is deleted from our live systems. If we restore a backup, we take reasonable steps to delete again anything that had already been deleted.
Actions logs 90 days.

Git keeps history: after an account is deleted, its commits can remain in other people's repositories, branches, tags, pull requests, forks, and clones.

Subprocessors

Our current subprocessor list is at https://legal.arkiviva.com/subprocessors/. Before a new subprocessor starts processing your data, we email you. If you object, you can end the affected service.

Where your data is

We keep customer personal data in the EEA: servers and backups in Germany and Finland, service email in France, and correspondence email in the Netherlands. If we need to process it outside the EEA, we tell you by email before we start, and we do so only with the safeguards the GDPR requires. Data you send elsewhere, for example through webhooks, mirrors, or your own runners, goes wherever you send it.

AI agents and training

Our staff, and the automated tools and AI agents they operate, help administer the service. We take reasonable steps to prevent AI agents from accessing customer personal data, except in public content: we do not use them to read content that is not public, such as private repositories, issues, wikis, and packages, unless we reasonably believe a task needs it; we do not copy customer personal data into prompts or work records; and we mask IP addresses in the server logs they read.

Our Terms do not allow us to use your content to train or fine-tune generative AI. We may use it to build tools that detect spam, malware, and abuse, as our Privacy Notice explains; those tools do not generate content.

Requests about personal data

If your request concerns personal data in a customer's content, please contact that customer, who decides about that data. We help customers answer such requests, and if you write to us at privacy@arkiviva.com, we can pass your request on. Requests about our own uses of personal data are covered in our Privacy Notice.