Privacy Policy — Helkrypt apps for Atlassian and customer support
Version 1.12 · Last updated: 18 November 2026 (1.12: Searchable Fields for Jira, section 4.9, qualifies for Runs on Atlassian; 1.11: added Searchable Fields for Jira, section 4.9; 1.10: Clear Transitions for Jira, section 4.6: the Rovo agent only lists the rules and cannot check or explain a single work item; 1.9: we handle support ourselves. Incoming mail to support@helkrypt.no is filtered by MailChannels, a copy is kept at Domene.no, and it is delivered to our inbox at Resend. We write and send the replies. No region is stated for those providers. The postal address and org. no. are stated; 1.8: Clear Tests for Jira, section 4.8, now describes the cleartests property and bug markers precisely; 1.7: added Clear Tests for Jira, section 4.8; 1.6: added Sprint Trail for Jira, section 4.7; 1.5: added Clear Transitions for Jira, section 4.6)
This policy explains how Helkrypt AI AS handles personal data in our apps for Atlassian cloud products and in our customer support. Our general privacy policy for the Helkrypt website and other services is at https://www.helkrypt.no/privacy.
1. Who we are
"We" means Helkrypt AI AS, a Norwegian company.
| Company | Helkrypt AI AS |
| Org. no. | 936351034 |
| Postal address | Postboks 14, 4756 Hovden |
| Privacy contact | support@helkrypt.no |
| Name on Atlassian Marketplace | Helkrypt |
2. What this policy covers
- Our apps on the Atlassian Marketplace (section 4)
- Support requests sent to support@helkrypt.no (section 3.1)
- Licence and sales information we receive from Atlassian (section 3.2)
3. Data we are responsible for (we are the controller)
3.1 Customer support
| Data | Your name and email address, the subject and text of your message, the file names of any attachments, email headers used to keep messages in the same thread, the case number and timestamps. We do not keep attachment contents: they are discarded when your email is processed. |
| Why | To answer your support request and keep a history of the case. |
| Legal basis | If you are a customer: performance of a contract (GDPR art. 6(1)(b)). If you are not a customer yet, for example with a question before buying: our legitimate interest in answering questions about our products (art. 6(1)(f)). |
| How long | Cases and messages are deleted 12 months after the case is closed. |
How it works: you email support@helkrypt.no. Incoming email is received and filtered by MailChannels, a copy is kept in a mailbox at Domene.no, and it is delivered to our support inbox at Resend. We write the replies and send them.
3.2 Licence information from Atlassian
| Data | Atlassian shares licence, trial and transaction information with us in its vendor portal. This can include the organisation name and the name and email of technical and billing contacts. This contact information stays in Atlassian's vendor portal: we do not copy it into our own systems. In our own reporting we store only aggregated figures from Atlassian's public Marketplace API, such as installs, ratings and reviews. |
| Why | Contact information: billing and support. Aggregated figures: understanding how our apps are doing. |
| Legal basis | Our legitimate interest in running our business (art. 6(1)(f)). Accounting records: legal obligation (art. 6(1)(c)). |
| How long | Contact information is kept by Atlassian under Atlassian's rules. Accounting records are kept as long as Norwegian accounting law requires. |
4. Data in your Atlassian site (we are a processor)
When your organisation uses one of our apps, your organisation is the controller for the data in its Atlassian site. We act only as a processor, and process that data only to run the app. The apps run on Atlassian's Forge platform.
All our current apps:
- keep all data inside Atlassian, in the same region as your site (data residency is supported). For Searchable Fields for Jira (4.9), data residency support has not been confirmed yet.
- send no data outside Atlassian, use no cookies and have no tracking or analytics
- qualify for Atlassian's "Runs on Atlassian" programme
- log only technical information (such as IDs and status codes), never work item content or names
- three apps, Release Studio for Jira, Clear Transitions for Jira and Clear Tests for Jira, use Atlassian's built-in AI (Forge LLMs), which runs inside Atlassian (see 4.5, 4.6 and 4.8)
4.1 Auto Issue & Description Templates for Jira
- Stores: template names and text, the projects and work types a template applies to, whether it is a default template, and when it was last changed. Stored in Forge hosted storage.
- Reads but does not store:a work item's description, project, work type, reporter and assignee. Names are used only to fill in template variables such as {{reporter}}.
- Personal data stored by the app:none. Template text can contain anything an administrator writes, so please don't put personal or confidential information in templates.
- How long: until an administrator deletes the template or the app is uninstalled. After uninstall, Atlassian deletes the app data under the Forge hosted storage lifecycle.
4.2 Epic Rollup for Jira
- Stores: admin settings (story points field per project, feature switches) in Forge hosted storage, and epic totals (number of child issues, story points, time, percent done, last updated) as a property on the epic. The totals contain only numbers and a date, never issue keys, titles or names.
- Reads but does not store:the epic's child issues (key, summary, status, story points, time fields, parent, Atlassian team).
- Optional: time logged per person.Off by default. If a Jira administrator turns it on, the app shows logged time per person on an epic. It is calculated in the viewer's browser, never stored, ranked or exported. Your organisation decides whether this is lawful (for example under employment law on monitoring staff) and must inform employees before turning it on.
- Personal data stored by the app: none.
- How long: settings until the app is uninstalled. Epic totals until an administrator turns off searchable totals (a background job then removes them) or the epic is deleted. Uninstalling does not remove the totals, because the app can no longer reach Jira after uninstall. Turn off searchable totals before uninstalling if you want them removed.
4.3 Smart Defaults for Jira
- Stores: rules in Forge hosted storage: rule name, the projects and work types it applies to, and the default values (priority, labels, components and number of days until the due date). The rules contain IDs, labels and numbers, not work item content.
- Reads but does not store:a new work item's project, work type, priority, labels, components and due date, to see which fields are empty. The app reads no user fields.
- Personal data stored by the app:none. Rule names and labels are free text written by an administrator, so please don't put personal or confidential information in them.
- How long: until an administrator deletes the rule or the app is uninstalled. After uninstall, Atlassian deletes the app data under the Forge hosted storage lifecycle. Values the app has set on work items are normal Jira field values and stay on the work items.
4.4 Comment & Reply Templates for Jira
- Stores: templates in Forge hosted storage: template name and text, the projects and work types it applies to, whether it is posted as an internal note or a reply to the customer in service projects, and when it was last changed.
- Reads but does not store:a work item's key, summary, project, work type, reporter and assignee, in the user's browser, to fill in template variables such as {{reporter}}. They only end up in the comment the user chooses to post.
- Comments: the user posts the comment as themselves. It is stored by Jira as a normal comment, not by the app.
- Personal data stored by the app:none. Template text can contain anything an administrator writes, so please don't put personal or confidential information in templates.
- How long: until an administrator deletes the template or the app is uninstalled. After uninstall, Atlassian deletes the app data under the Forge hosted storage lifecycle. Posted comments are normal Jira comments and stay on the work items.
4.5 Release Studio for Jira
- Stores: templates, rules, the list of internal words and settings (whether AI is allowed, how long release records are kept) in Forge hosted storage. Also release records: when a user chooses "Freeze release record", the app saves the release notes text, the project, version names, template name, the work item keys and their status, the time, and a checksum. A record can't be edited afterwards. Everyone who can see the project sees the list of records (versions, template, date and status), but only project administrators, Jira administrators and the person who froze a record can open its text and export it. If a template requires approval, release notes made with it can only be published to Confluence from an approved record.
- People in records:each record keeps the Atlassian account ID of the person who froze it and, if approval is used, of the person who approved it. The app doesn't store or show names. Every week the app reports these account IDs to Atlassian's personal data reporting service, and removes the account ID from a record when Atlassian reports that the account is closed.
- Reads but does not store:work items (key, summary, status, work type, labels, components, version, and an optional release note field) to build the release notes in the user's browser, and the list of Confluence spaces so the user can choose where to publish.
- AI:only when a user selects "Improve with AI" or "Translate" (or an administrator's automation rule or Rovo agent does so), the app sends the work item key, its text (at most 1,500 characters), work type, labels and components to Atlassian's built-in AI (Forge LLMs). The AI runs inside Atlassian. The data doesn't go to us or to any other company, and the app sends no names, reporters, assignees or links. Administrators can turn AI off for the whole site or for a single template. The AI result is kept briefly in Forge hosted storage so the page, the Rovo agent or the automation rule can use it: the Releases page deletes it once fetched, and a daily cleanup removes every AI result older than one hour, so none is kept longer than about 25 hours. For the Rovo agent, the stored AI result also holds the account ID of the person who asked for it, so only that person can publish it.
- Confluence:when a user selects "Publish to Confluence", the app creates a normal Confluence page as that user. The page is stored by Confluence, not by the app, and stays after uninstall.
- Logs: only the type of operation, IDs, status or error codes and AI token counts. Never work item content, AI text or names.
- Personal data stored by the app:account IDs in release records (see above). Work item titles, template text and release notes can contain anything your teams write, so please don't put personal or confidential information in them.
- How long: templates and settings until an administrator deletes them or the app is uninstalled. Release records for the retention period your administrator sets (default 7 years, between 1 and 25 years), then they are deleted automatically, or until the app is uninstalled. After uninstall, Atlassian deletes the app data under the Forge hosted storage lifecycle.
4.6 Clear Transitions for Jira
- Stores: your rules (name, type, settings and the message shown to users), a list of where each rule is used (workflow, transition and project IDs and the rule version), and one setting (how long the block log is kept) in Forge hosted storage. A copy of each rule is also saved in your Jira workflow by Jira, and it stays there until an administrator removes the rule from the transition.
- Block log (optional per rule):if an administrator turns on "Log blocks" for a rule, the app saves the work item key, the rule ID, the transition ID and the time each time a transition is blocked. The log contains no Atlassian account IDs and no names. It is kept for 90 days by default (an administrator can choose between 30 and 365 days), then deleted automatically, or until the app is uninstalled.
- Reads but does not store:work items (status, fields, links and child work items) when Jira checks a rule and when an administrator tests a rule. These checks run as the signed-in user, or inside Jira for the rule itself. The optional "(logged)" rule type is different: it runs app code that evaluates the rule as the app, inside Atlassian, to write the block log. It still stores no person data. The app doesn't save the results.
- AI:only when an administrator selects "Draft with AI", the app sends the administrator's text (at most 500 characters) and the names and IDs of fields, work types and link types to Atlassian's built-in AI (Forge LLMs). The AI runs inside Atlassian. No work item content and no names of people are sent. The app doesn't save the text.
- Logs: only fixed technical keys and counts. Never work item content, rule messages, AI text or names.
- Personal data stored by the app:none by design. Rule names and messages are free text that your administrators write, so please don't put personal information in them.
- Uninstalling: Atlassian deletes the app data under the Forge hosted storage lifecycle. Rules saved in your workflows are not removed by uninstalling and keep blocking transitions until an administrator removes them, so remove them first.
4.7 Sprint Trail for Jira
- Stores: on each work item that has sprint history, a small property called
sprintTrail: the IDs of the sprints the work item has been in, added to late, removed from, carried over from and completed in, two counts, and the time it was last updated. It contains no text and no names. The property is stored in Jira, with the work item. In Forge hosted storage the app keeps sprint information (sprint ID, name, state, dates and board ID), the administrator's settings (which status categories count as done and which link type means blocked) and the status of the history build. - Reads but does not store:a work item's sprint and status changes, links, flags and estimate, and its key, summary and status, to work out the reasons work spilled over and to show them to the user. The app does not read or store who made a change. Views and Rovo actions read as the signed-in user, so they only show work items that user can see. Sprint names used in searches are looked up as the app and turned into sprint IDs.
- AI:the app has no AI service of its own and sends no data to Atlassian's built-in AI (Forge LLMs). The Sprint Trail Rovo agent is part of your Atlassian site: when a user asks it a question, the app's read-only actions return sprint facts (work item keys, summaries, statuses and reasons) to your Rovo, which explains them. The data doesn't go to us or to any other company.
- Logs: only the type of operation, the event type, work item and sprint IDs, counts and fixed error keys. Never work item content, sprint names, names of people or error text from Jira.
- Personal data stored by the app:none by design. No Atlassian account IDs and no names are stored. Sprint names are free text that your teams write, so please don't put personal information in them.
- How long: the
sprintTrailproperty stays as long as the work item exists and the app is installed. The app removes it from all work items when you uninstall (a pre-uninstall step asks Jira to delete it, which Jira completes in the background), and an admin can remove it earlier with Remove sprint history. It holds only sprint IDs, sprint dates (as days) and counts. The data in Forge hosted storage is kept until the app is uninstalled, and Atlassian then deletes it under the Forge hosted storage lifecycle.
4.8 Clear Tests for Jira
- Stores in Forge SQL (one database per installation): your tests (title, objective, preconditions, steps, Gherkin text, parameter tables, labels, automation key), folders, the version history of each test, links between tests and work item IDs, test cycles, results per step, actual results, comments, and CI runs with the result log (status, duration and a CI message of at most 2 KB per result). Tests you delete stay in a trash for 30 days.
- People in the data:each row keeps the Atlassian account ID of the person who created or changed it, was assigned a test, or ran it. The app doesn't store or show names or avatars: Jira shows them when a page is displayed. A cycle export (CSV) contains these account IDs. The app is not made to measure or rank employees, and it shows no statistics per person apart from "assigned to me" and progress within one cycle.
- Stores in Forge hosted storage: settings (site and project), CI tokens (only a one-way SHA-256 hash with a name and dates; the token itself is shown once), a monthly AI counter and a marker for the personal data report. Short-lived records exist while you use AI or import a file: AI jobs (with the account ID of the person who started them and the drafts), uploaded file text, import plans and import status. They expire after 24 hours and are removed automatically, normally within a day and at the latest within about two days. When a bulk action creates a Jira bug that the app cannot store on a test result, it keeps a marker with the bug ID and key for 30 days. The marker holds no personal data.
- Stores in Jira: on each work item that has or had linked tests, a small property called
cleartestswith the number of tests, how many passed, failed and not run, the status and the time it was updated. It contains no text and no names. When no tests are linked any more, the counts go to zero and the property stays. - CI import: your own CI pipeline sends test reports (JUnit XML or Cucumber JSON) to a web address hosted by Atlassian, with a token that belongs to one project. The app checks the token, stores the results in that project and never creates Jira work items from a report. Nothing is sent from the app to the outside.
- AI:only when a user selects an AI action, the app sends text to Atlassian's built-in AI (Forge LLMs), which runs inside Atlassian. Generating tests and finding gaps send the work item's summary, description and acceptance criteria and the titles of linked tests (finding gaps also the steps of linked tests). Quality check and conversion to or from Gherkin send the text of that one test. Summarizing failures sends the titles of failed tests, their expected and actual results and the CI failure messages. No people fields (reporter, assignee, names, account IDs) are sent. The data doesn't go to us or to any other company. AI only makes drafts: nothing is saved until a person selects Accept. A Jira administrator can turn AI off or set a monthly limit. The Clear Tests Rovo agent uses your organisation's own Rovo. An optional MCP server for coding assistants is off by default and an administrator turns it on.
- Logs: only the type of operation, IDs, counts, HTTP status codes and fixed error keys. Never test text, work item text, tokens or account IDs.
- Personal data stored by the app:account IDs (see above). Tests, comments, results and CI messages are free text that your teams write, so please don't put personal information in them.
- Reporting:every week the app reports the stored account IDs to Atlassian's personal data reporting service and clears the account ID in the data when Atlassian reports that the account is closed.
- How long: tests, cycles and results until you delete them or the app is uninstalled; deleted tests leave the trash after 30 days. An administrator can set how long CI run history and the CI result log are kept (default: keep everything); the latest CI result of each test, with its message, stays on its cycle until the cycle is deleted. The
cleartestsproperty stays on every work item that has or had linked tests, with zero counts once none are left, until the app is uninstalled or an administrator uses Remove coverage data. The app removes it from all work items when you uninstall (a pre-uninstall step asks Jira to delete it, which Jira completes in the background), and an administrator can remove it earlier. After uninstall, Atlassian deletes the data in Forge hosted storage and Forge SQL under the Forge lifecycle. If the license is not active, the app is read-only: you can still view and export your data, and nothing is deleted.
4.9 Searchable Fields for Jira
- Stores: on each work item, five read-only custom fields: Last comment date (date and time), Last comment author (the Atlassian account ID of the author), Comment count, Subtask count and Open subtask count. The values are stored in Jira, with the work item. Comment text is not stored. In Forge hosted storage the app keeps only the status of the background build (state, how many work items have been processed, an approximate total, a page token, timestamps, a run ID and a fixed error key) and a cache of the five fields' Jira IDs.
- Reads but does not store: comments (creation time, author and whether the comment has a visibility restriction) and subtasks (whether the work item is a subtask, its status category and its parent), to compute the field values. These reads and the field writes run as the app, inside Atlassian. Jira includes comment text in the data the app fetches, but the app does not use or store it. Comments restricted to a role or group are not counted and never become the last comment. In Jira Service Management, internal comments are counted, because Jira does not mark them with a visibility restriction. The app uses three scopes:
read:jira-work,write:app-data:jiraandstorage:app. It has no admin scopes, no remotes and no egress. - AI: the app has no AI service of its own and sends no data to Atlassian's built-in AI (Forge LLMs).
- Logs: only the type of operation, counts and fixed error keys. Never comment text, names or account IDs.
- Personal data stored by the app: the Atlassian account ID in the Last comment author field, on the work item. The app does not store names; Jira shows the name when the field is displayed. The account ID is not copied into Forge hosted storage, and the app has no process of its own to clear the field if the account is later closed.
- How long: the five field values stay on the work item while the app is installed. The app does not remove them when you uninstall (it has no uninstall step and no action to remove the values). According to Atlassian's lifecycle for Forge custom fields, the fields disappear from Jira when the app is uninstalled. If you reinstall within 30 days, the fields and their values come back. After 30 days, the field data is permanently lost. The data in Forge hosted storage is kept until the app is uninstalled, and Atlassian then deletes it under the Forge hosted storage lifecycle. No data is stored on Helkrypt's own servers.
4.10 Requests about data in your site
Send requests about access, correction or deletion of data in your Atlassian site to your own site administrator. We help your organisation answer them. Customers who need a data processing agreement can find ours at https://www.helkrypt.no/dpa.
5. Customer support
We handle customer support ourselves. A person at Helkrypt reads your message and writes the reply. Nothing is decided about you by software alone, so there is no automated decision-making under GDPR art. 22.
Three of our apps, Release Studio for Jira, Clear Transitions for Jira and Clear Tests for Jira, can use Atlassian's built-in AI when a user chooses to. See sections 4.5, 4.6 and 4.8.
6. Service providers and international transfers
We use these providers. All of them have signed a data processing agreement with us.
| Provider | What they do for us | Where data is processed | Transfer safeguard |
|---|---|---|---|
| Atlassian | Hosts our apps (Forge), Marketplace and licensing | Your site's region; Atlassian also operates in the US | EU–US Data Privacy Framework and Standard Contractual Clauses (Atlassian DPA) |
| Hostinger | Hosts our workflow automation server | Germany (Frankfurt) | Within the EEA (Hostinger DPA) |
| MailChannels | Receives and filters incoming email for support@helkrypt.no. Public MX records: mx1.mailchannels.net and mx2.mailchannels.net. | Not stated | Not stated |
| Domene.no | Our domain and email provider. Keeps a copy of support email in a mailbox. | Not stated | Not stated |
| Resend | Holds our support inbox at support@helkrypt.no. We write the replies, and Resend sends them. | Not stated | Data processing agreement (Resend DPA) |
7. How long we keep data
| Data | Retention |
|---|---|
| Support cases and messages | 12 months after the case is closed |
| Support email at Resend | 30 days |
| Licence contact information | Stays in Atlassian's vendor portal, under Atlassian's rules |
| Aggregated Marketplace figures | While useful for running the business |
| Accounting records | As required by Norwegian accounting law |
| Data in the apps | See section 4 |
8. How we protect data
- Access is limited to the people and systems that need it.
- Passwords and API keys are kept in secret stores, never in code.
- Data is encrypted in transit (TLS).
- We collect as little as possible: attachment contents are discarded, successful automation runs are not stored, licence contact information stays with Atlassian, and our apps store no personal data of their own, apart from the account IDs in Release Studio for Jira's release records (see 4.5), the account IDs and customer-written content in Clear Tests for Jira (see 4.8), and the account ID in Searchable Fields for Jira's Last comment author field on the work item (see 4.9).
9. Your rights
You have the right to:
- access the personal data we hold about you
- correct inaccurate data
- delete your data
- restrict how we use it
- receive your data in a portable format
- object to processing based on our legitimate interest
Email support@helkrypt.noto use your rights. We answer within one month. For data in your organisation's Atlassian site, see section 4.10.
10. Complaints
You can complain to the Norwegian Data Protection Authority, Datatilsynet: https://www.datatilsynet.no/en/about-us/contact-us/how-to-complain-to-the-norwegian-dpa. You can also complain to the authority in the EEA country where you live or work.
11. Changes
When we change this policy, we update the version number and the "last updated" date at the top.