Key takeaways
- Jira shows your work item total as the result count of a JQL search under Filters > View all work items, but it counts only the projects the searcher can browse.
- A Jira admin exports the whole site from Settings > System > Backup manager and ticks the option to include attachments, avatars and logos.
- A backup that includes attachments can only be made every 48 hours, and a CSV export covers up to 10,000 work items, so split larger sets by project or date.
- You don't need to export anything to apply: have your work item count, years of history and seat count ready, and the data vendor anonymizes the records before any AI lab sees them.
01
What is Jira data, and why do AI labs and data vendors buy it?
The short answer
Jira data is the history of a company's tracked work: every work item (a task, bug or request), its status changes, the comments on it and the links to related items. AI labs (companies that build and train AI models) and data vendors (companies that anonymize business records and package them for those labs) buy a copy because a work item follows one piece of work from the day it was raised to the day it was finished.
What a Jira site records
Jira is Atlassian's app for tracking work. A company's site is its own Jira, and its records are organized in a few layers.
- Project
- A collection of work for one team or product, with its own permissions.
- Work item
- One piece of work, such as a bug to fix or a feature to build. Older Jira documentation calls it an issue.
- Status and workflow
- The stages a work item moves through, such as To Do, In Progress and Done, and the rules for moving between them.
- Sprint
- A fixed period, usually one to four weeks, in which a team commits to finish a set of work items.
- Comment and link
- The discussion between the people doing the work, and the connections between related work items.
Why a work item teaches an AI model something
An AI model learns by studying a very large number of examples and finding the patterns in them, and those examples are called training data. A work item is a complete example of a job: the description of the problem, each decision along the way, the blockers and the final resolution, all in order.
AI labs now want to train AI agents, software that carries out a job step by step. An agent can learn how a team breaks a large request into smaller tasks by studying thousands of work items. Estimates, sprints and custom fields add structure, because they allow a dataset to link the way work was planned to the time it really took and the obstacles it met. The full mechanism is in AI training data: why AI companies want your company's data.
How Jira fits into a dataset
Jira is one of several apps in a company's dataset, which usually also includes chat, email and documents. Companies are paid $100K to $2M or more per dataset, according to micro1, the data vendor we typically refer companies to. The price covers the whole dataset, not each app on its own, and what your data is worth explains what moves it.
02
Where does Jira show how much data you have?
The short answer
Jira shows the total number of work items as the result count of a search. Open Filters > View all work items, switch to JQL (Jira's search language), search with ORDER BY created ASC and read the total. The count works on every Jira Cloud plan, but it includes only the projects the searching person can browse.
How to count your work items
The steps below read the size of a Jira site from its own search. They use only the work item navigator, the page where Jira lists search results.
- Log in as someone who can browse every project. The search counts only the work items its user is allowed to see.
- Open Filters and choose View all work items.
- Switch the search box to JQL, enter just ORDER BY created ASC (no filters), and press search.
- Note the total that the navigator reports. Because the list runs oldest first, the top result is your earliest work item.
Who can see the full count
Search exists on every Jira Cloud plan, so no plan upgrade is needed. What changes the number is permission: a Jira admin may have to grant browse access to every project first. If your company also runs Jira Service Management (Atlassian's help-desk product) on the same site, count it as its own system.
Why comments and attachments do not add to the count
Comments, attachments and history belong to each work item, so one work item with fifty comments is still one record in the count. The total is therefore the number of pieces of work, not the number of words or files. The size of the files shows up later, in the backup.
03
What should you know about your Jira data before you apply?
The short answer
Have three facts ready: how many work items your Jira site holds, how many years it has been in use and how many people hold a license. They go into the data manifest, the written summary of which apps hold your records, how many years each covers and roughly how much. Applying shares no data, and you don't need to run an export first.
The number of work items
Report the total from the all-projects search as records. If the person searching could not browse every project, say so, because the figure is then a minimum.
The years of history
Look at the creation year of the top result in that same oldest-first search, and it tells you how long Jira has been in use. If the company moved from another tracker and imported old tickets, mention the move.
The number of seats
Seats are the licensed users on your Jira subscription. A Jira Service Management subscription on the same site is counted separately.
What you don't need to do first
You don't need to export anything to apply, and you don't need to clean or scrub the work items. The data vendor anonymizes the records before anyone else sees them, and nothing moves until a lawyer-reviewed license is signed. The application, the intro call and each later step are laid out in how to sell your company's data.
04
How do you export your full Jira history?
The short answer
A Jira admin creates a site backup from Settings > System > Backup manager, ticks the box to include attachments, avatars and logos, and downloads the zip file. For a spreadsheet of one slice of work items, a JQL search can be exported as CSV, up to 10,000 work items at a time.
Who can export
Run the backup as a Jira admin. Atlassian notes that an organization can choose to prevent exports, so if the option is missing, check with your organization admin before assuming something is broken.
The steps for a full site backup
- Log in as a Jira admin.
- Go to Settings > System and, under IMPORT AND EXPORT, select Backup manager.
- Under Backup for cloud, select Create backup for cloud, and tick the option that includes attachments, avatars and logos.
- When the backup is ready, select Download backup file. The zip contains entities.xml, activeobjects.xml and a data folder holding the attachments.
These steps follow Atlassian's export guide. Atlassian suggests keeping a backup within about 25 GB of data, and 50 to 100 GB of attachments.
The limits that apply to backups
A backup that includes attachments, avatars or logos needs 48 hours before the next one can be made. Since January 22, 2026, Backup manager restores only backups that are 30 days old or less, so download the file promptly and store it yourself.
Premium and Enterprise sites can also use Atlassian Backup and Restore at admin.atlassian.com. It needs both organization admin and app admin permissions, and it includes attachments.
How to export a slice as a spreadsheet
- Run a JQL search in the work item navigator for the work items you want.
- Open More actions > Export and pick Export CSV - all fields. In the file, every comment sits in a separate Comment column.
- Wait for Jira's notification, then follow its link to download the file.
Atlassian's guide to exporting over 10,000 work items says the asynchronous CSV export handles up to 10,000 work items. For a larger set, split the search by project or created date, export each part and merge the files, then check that the combined count matches the original search.
What each export holds
The table compares the two routes, and the diagram after it shows the same site exported both ways.
| Export | Size | Attachments | Best for |
|---|---|---|---|
| Backup manager | Whole site, ideally up to about 25 GB of data | Files, if ticked | The complete history |
| CSV export | Up to 10,000 work items per export | URLs only, not files | One project or date range |
What a backup leaves out
A backup covers the work items with their built-in and custom fields, the comments, the boards and sprints, user settings and each project's configuration (workflows, schemes and screens). Three things are missing: automation rules (exportable on their own), data from third-party apps, and Assets, the Jira Service Management feature that tracks equipment and other business objects.
How much storage each plan holds
Attachments make up most of a large Jira site's size. Atlassian's attachment guide lists 2 GB per app on Free, up to 250 GB on Standard and unlimited on Premium. The default maximum for one file is 1 GB, and an admin can raise it to 2 GB.
How long an export takes
Atlassian gives no time estimate for a backup, and a CSV export of a large search can take several minutes. Start early and wait for Jira's notification.
Sources: Atlassian Support, Export data from Jira Cloud, Export over 10,000 work items in Jira Cloud and Configure file attachments. Steps checked against these pages in November 2026.
05
What stays private when Jira data is licensed?
The short answer
The company decides which projects are in scope, and the data vendor replaces names and other personal details with consistent stand-ins before an AI lab sees anything. The company keeps ownership of its Jira history, and the license covers defined records for a defined use.
You choose what is in scope
A company can leave security, HR or customer-specific projects out of what it licenses. The data vendor also deletes records that are too sensitive to clean safely instead of anonymizing them.
Names are replaced before a lab sees them
Anonymization means replacing every detail that identifies a person, such as a name or an email address, with a stand-in that is the same every time it appears. The work item still reads in order, so a lab can learn from it, but it no longer says who was involved. The diagram shows one work item before and after.
The method, including what is deleted outright and how you review samples, is in how company data is anonymized for AI training.
Ownership and payment
Licensing gives the buyer limited rights to a copy, while the company keeps ownership of the originals. Exclusivity, permitted use and deletion are covered in data licensing. micro1 pays in full within 30 days of accepting the data, usually as a single payment.
Questions
Frequently asked questions
Why does our Jira CSV export stop short of all our work items?
A CSV export from the work item navigator handles up to 10,000 work items at a time, and Atlassian's older single-request method was capped at 1,000. Beyond that, split the search by project or created date and merge the files, or use Backup manager for the whole site.
Does the Jira Cloud backup include attachments?
Yes, if you tick the option that includes attachments, avatars and logos when you create the backup. A backup that includes them can be made only once in any 48-hour period.
Who is allowed to back up a Jira Cloud site?
A Jira admin who can open the System settings. If an organization admin has set a data protection policy that blocks exports, the option does not appear.
How much storage does each Jira plan allow?
Storage is 2 GB per app on Free, up to 250 GB on Standard and unlimited on Premium. A single file can be 1 GB by default, and an admin can raise that cap to 2 GB.
Do we need to export Jira before we apply?
No. The application takes about two minutes and shares no data. For the intro call you need your work item count, your years of history and your seat count, and you can read the first two from a single JQL search. The full export is needed later, when the records are transferred after you sign the license.
Can we license only some Jira projects?
Yes. The company chooses the projects in scope, so security, HR or customer-specific projects can stay out. Because the work item count covers only projects the searcher can browse, make sure your count matches the projects you intend to license.
What is Jira data worth to AI labs and data vendors?
Jira history is priced as part of a whole dataset, not on its own. Companies are paid $100K to $2M or more per dataset, according to micro1, and the price depends on the years of history and how completely the records show each job. What your data is worth explains the factors.
Bottom line
Know what is in Jira, then see what it is worth
You don't need to export anything to apply. Apply in two minutes, describe your records on the intro call, and micro1, the data vendor we typically refer companies to, prices the whole dataset and sends an offer 1–2 days later.
Not legal, tax or financial advice. Figures come from micro1's published pages and public reporting as of November 2026.