Which Atlassian apps are making real money now?

Is the Atlassian Marketplace actually big enough to build a serious software company?

Yes. The Atlassian Marketplace is currently large enough to support businesses ranging from profitable bootstrapped companies to software groups generating well above $100 million in recurring revenue.

Atlassian says Marketplace lifetime sales have now passed $6 billion, across more than 6,000 apps and integrations from over 2,000 partners. The more revealing figure is the speed of accumulation. The Marketplace crossed $4 billion only in early 2024; Atlassian subsequently reported $5 billion during 2024 and more than $6 billion by late 2025. In other words, roughly $2 billion of the Marketplace's first $6 billion arrived in less than two years.

This is no longer an ecosystem where $1 million of lifetime revenue makes a vendor exceptional. Atlassian was already talking about 50 Marketplace businesses that had grossed more than $1 million when lifetime Marketplace sales were only around $1 billion.

The underlying customer pool has expanded too. Atlassian now serves more than 300,000 customers, including a large majority of the Fortune 500, while its own annual revenue has passed $6.5 billion. Marketplace vendors can therefore sell relatively narrow enterprise functionality into a very large installed software base without building their own distribution system from scratch.

Marketplace measure	Scale now
Lifetime Marketplace sales	$6B+
Apps and integrations	6,000+
Marketplace partners	2,000+
Atlassian customers	300,000+

Why is it so hard to know which Atlassian apps make the most money?

The highest-grossing Atlassian apps are surprisingly difficult to rank because Atlassian does not publish revenue by app, and the companies behind the biggest products usually keep those numbers private.

Marketplace pages expose installs, reviews, prices and vendor identities, but not gross sales. Even installs need interpretation. An installation can belong to a ten-person company or an enterprise with thousands of licensed Jira users, while Atlassian apps are generally priced according to the number of users on the underlying Jira or Confluence instance. Two apps with identical install counts can therefore have dramatically different revenue.

Acquisitions further blur the picture. BigPicture now belongs to Appfire, Structure belongs to Tempo, and Zephyr belongs to SmartBear. Appfire itself owns more than 100 applications across several software ecosystems. The Adaptavist Group combines Marketplace software such as ScriptRunner with consulting and Atlassian resale. Their company revenue cannot simply be assigned to one plugin.

So we need several tests at once: disclosed app ARR when available, customer counts, Marketplace installs, vendor financial disclosures, acquisitions, enterprise adoption and evidence that customers continue paying after the transition from Server to Cloud and Data Center. That gives a much better picture than sorting Marketplace results by installs.

Which Atlassian app has the clearest proof of making serious money?

Structure is one of the strongest publicly verifiable examples because its economics were disclosed: the company behind Structure was already generating roughly $23 million of annual recurring revenue with EBITDA margins above 20% when Tempo acquired it.

That disclosure came from listed Icelandic technology company Origo when Tempo bought ALM Works, Structure's developer, in 2021. ALM Works also reported more than 5,000 customer organizations at the time.

Structure has not disappeared since the acquisition. The current Structure PPM listing has roughly 13,000 Marketplace installs, and Tempo has incorporated it into a broader strategic portfolio-management product line alongside Timesheets, Capacity Planner, Roadmunk and other acquired products.

The important observation is the order of magnitude. Structure did not need 50,000 customers to become a $20-million-plus recurring-revenue business. Enterprise Jira installations are large enough that a specialized PPM product can generate substantial revenue from a few thousand organizations.

If we are looking for evidence that an individual Atlassian app can become a genuine mid-sized SaaS company rather than a profitable side project, Structure provides perhaps the cleanest public proof.

Is Tempo Timesheets still one of the biggest Atlassian apps?

Yes. Tempo Timesheets remains one of the clearest Atlassian Marketplace cash generators, even though Tempo does not disclose Timesheets revenue separately.

Timesheets currently has roughly 27,000 Marketplace installs and has been sold in the Jira ecosystem since 2009. More importantly, Tempo says its overall product portfolio serves more than 30,000 organizations, including approximately one-third of the Fortune 500. Atlassian itself describes Timesheets as the most popular time-tracking app for Jira.

Tempo has built an organization of hundreds of employees around this customer base and repeatedly expanded through acquisitions. It bought ALM Works, whose Structure business alone was already producing around $23 million ARR, as well as Roadmunk, Prime Timesheet, LiquidPlanner, Old Street Solutions and Alpha Serve.

That acquisition pattern tells us something installs alone do not. Tempo is not treating Jira time tracking as a small plugin category. It is using the cash flow and customer relationships generated around products such as Timesheets to assemble a broader portfolio-management platform.

Timesheets may no longer represent the majority of Tempo's future strategy, but commercially it belongs in the first group of Atlassian apps we can confidently describe as substantial software businesses.

Is ScriptRunner for Jira making real money or just benefiting from its age?

ScriptRunner is almost certainly one of the largest Atlassian apps commercially, and its current footprint is too large to dismiss as historical inertia.

ScriptRunner for Jira has around 35,000 active Marketplace installs, making it one of the most widely installed paid Jira extensions we found. Its appeal is unusually deep: administrators use it for scripting, workflow customization, automation, custom JQL and functionality that can become embedded in how an organization operates Jira.

The vendor behind it also has unusual scale. The Adaptavist Group recently reported £325 million in annual revenue, up 7%, although we have to be precise about what this tells us. Adaptavist also sells professional services, cloud migrations, training and software licenses from partners including Atlassian, AWS and GitLab. £325 million is not ScriptRunner revenue.

What the figure does prove is that ScriptRunner sits inside one of the largest companies ever built around the Atlassian ecosystem. Its 35,000-install installed base is larger than Xray, Tempo Timesheets, JMWE, BigPicture and Structure individually.

We cannot responsibly attach a current ARR number to ScriptRunner. We can, however, put it firmly among the handful of Atlassian apps whose commercial scale is no longer in doubt.

Is Xray Test Management for Jira a major software business?

Yes. Xray is one of the strongest candidates for a large eight-figure Atlassian app business, although its owner does not publicly disclose Xray's revenue.

Xray currently has roughly 25,000 Marketplace installs. That places it among the largest specialized paid Jira apps, despite operating in a narrower category than general workflow automation or diagramming.

Its economics are helped by the nature of the customer. Test-management software is not generally installed as a cosmetic Jira enhancement. QA, engineering and compliance processes can depend on test cases, traceability, automated-test integrations and release evidence. Once these workflows are built around a testing system, replacement becomes operationally expensive.

Xray's owner, Xblend, has consequently expanded the product upward with Xray Enterprise rather than competing only on a cheap plugin. It also sells Xporter, which has several thousand additional installs. Atlassian recognized Xray for Marketplace channel growth in its 2024–2025 partner awards.

The comparison with Zephyr makes the category even more convincing. Two different test-management products have each sustained five-figure Jira install bases for years. That suggests we are looking at one of the Marketplace's most economically attractive software categories rather than one unusually successful app.

Jira testing app	Current Marketplace installs
Xray Test Management	~25,000
Zephyr Essential	~10,000
Xray Enterprise	~1,200

Is draw.io quietly one of the biggest Atlassian Marketplace businesses?

Yes. draw.io is probably one of the largest paid Atlassian apps by distribution, with more than 60,000 active Confluence installs.

The current draw.io Confluence listing shows roughly 62,000 installs, while its separate Jira app adds close to another 9,000. Few paid Marketplace products come close to that footprint.

Its business model is different from Xray or Structure. Diagramming is much broader but generally lower-value per user than enterprise test management or portfolio management. This makes raw install comparisons misleading: 60,000 draw.io installations do not automatically imply three times the revenue of a 20,000-install enterprise app.

Yet the breadth is economically powerful. Confluence pages naturally accumulate architecture diagrams, process maps, UML, network drawings and business workflows. Once those documents are embedded throughout a company's knowledge base, the app becomes part of the organization's stored knowledge rather than a disposable productivity tool.

draw.io has also been one of the earlier large Marketplace vendors to move major functionality onto Forge. That is increasingly important because Atlassian is financially favoring apps that fully adopt its newer platform.

We do not have a public draw.io ARR figure, but by installed paid distribution it clearly belongs in the top tier.

How much money are Appfire's Atlassian apps actually making?

Appfire proves that Atlassian apps can support a software company with more than $200 million in recurring revenue, but we should not pretend that the entire $200 million comes from Jira and Confluence.

Appfire disclosed $200 million ARR in early 2024 after growing from approximately $10 million ARR in four years. It subsequently said it had moved well beyond $200 million ARR while remaining profitable. The company employs more than 800 people and has assembled a portfolio exceeding 100 apps.

Atlassian remains central to that portfolio. Appfire owns BigPicture, JMWE, JSU Automation Suite, Rich Filters, Comala Document Management, Configuration Manager for Jira, Power Scripts, BigGantt and numerous other established Jira and Confluence products.

Several have substantial individual footprints. JMWE has roughly 16,800 installs, JSU about 12,000, BigPicture nearly 11,000, Rich Filters around 9,500 and Comala Document Management more than 6,000.

Appfire has nevertheless expanded into monday.com, Microsoft, Salesforce and other ecosystems, so its company ARR cannot be used as an Atlassian revenue figure.

The useful conclusion is different: Marketplace apps have produced enough recurring cash flow and acquisition opportunities for someone to assemble a profitable $200-million-plus portfolio around them. That puts the upper end of the Atlassian app economy far beyond the normal definition of an indie-plugin business.

Are Jira workflow apps still lucrative now that Jira has its own automation?

Yes. Advanced Jira workflow apps are still making substantial money because Atlassian's native automation has reduced simple use cases without eliminating complex ones.

ScriptRunner remains around 35,000 installs. Appfire's JMWE has about 16,800, while JSU Automation Suite has roughly 12,000. Jira Workflow Toolbox remains near 5,000. These are not abandoned legacy apps waiting for native functionality to kill them.

The reason becomes clearer when we compare what companies actually buy. Native Jira automation handles many trigger-and-action workflows well. ScriptRunner, JMWE and similar products increasingly sell deeper administration capabilities: scripted conditions, custom searches, field behavior, sophisticated validators, integrations and transformations across large Jira environments.

Atlassian has commoditized the bottom of the automation market while leaving room at the top.

There is a warning here for new developers. A new app whose entire proposition is “one extra Jira automation action” faces much worse economics today. A product that becomes the administration layer for hundreds of workflows is different. The continuing scale of ScriptRunner, JMWE and JSU shows that customers still pay for that distinction.

Why do project and portfolio-management apps make so much money on Atlassian?

Project and portfolio management is one of the most lucrative Atlassian app categories because vendors can sell executive-level planning on top of Jira data without replacing Jira itself.

Structure reached approximately $23 million ARR before its acquisition. BigPicture currently has around 11,000 Marketplace installs. Tempo subsequently combined Structure with Timesheets, capacity planning, financial management and roadmapping to move further into strategic portfolio management.

This category has unusually attractive economics. The underlying Jira deployment may include hundreds or thousands of employees, while the buying problem is valuable: resource allocation, cross-team dependencies, budgets, programs, capacity and executive visibility.

That allows a Marketplace company to capture a small portion of the value of an enterprise PPM system while Atlassian provides much of the underlying work-management infrastructure.

It also explains why acquisition activity has been so intense. Tempo bought ALM Works; Appfire bought SoftwarePlant, the developer of BigPicture. SoftwarePlant already employed about 165 people and reported more than 26,000 BigPicture-related installs when Appfire bought it in 2021.

The biggest Atlassian businesses therefore tend not to solve tiny interface annoyances. They increasingly sit one management layer above Jira and turn operational data into planning, financial or governance decisions.

Does test management look more durable than most Jira app categories?

Yes. Test management appears unusually durable because Jira does not eliminate the need for a dedicated testing system simply by improving its core issue tracker.

Xray has around 25,000 installs and Zephyr Essential around 10,000. Historically, Zephyr's economics became significant extremely quickly: Atlassian's own developer case study reported more than $100,000 of revenue during Zephyr's first quarter and nearly 10,000 customers by 2016.

The category has subsequently matured rather than disappeared. Both Xray and Zephyr support Cloud and Data Center, automation frameworks and enterprise workflows. Xray has added an Enterprise edition, while SmartBear has continued developing Zephyr inside a much larger software-testing portfolio.

That persistence is important. Many Marketplace apps are essentially feature-gap businesses: Atlassian can eventually absorb the feature. Test management is closer to a vertical application built on Jira. Customers need test libraries, execution histories, coverage, traceability and connections to CI/CD systems.

The difference helps explain why multiple vendors can maintain substantial businesses in the same category instead of one Marketplace bestseller absorbing the whole market.

Are Confluence apps making serious money too, or is Jira where all the money is?

Confluence supports serious Marketplace businesses, but the winning products tend to turn Confluence into something more specialized rather than merely add formatting.

draw.io is the obvious giant with more than 60,000 installs. K15t has built an entire documentation software business around its Scroll products: Scroll PDF Exporter alone has close to 9,000 installs, Scroll Sites more than 3,000 and Scroll Content Manager roughly 2,600.

Appfire's Comala Document Management has more than 6,000 installs and adds formal document workflows and governance to Confluence. These customers are buying controlled publishing, documentation management, approvals or specialized output—not another color picker.

The pattern is similar to Jira. draw.io turns Confluence into a diagramming environment. Scroll turns it into a technical-documentation platform. Comala turns it into a controlled document-management workflow.

That specialization supports higher willingness to pay and makes the app harder to replace with a small native feature.

Confluence therefore has fewer obvious mega-categories than Jira, but it has produced several durable software companies.

Can a newer cloud-only Atlassian app still become a real business?

Yes. Newer cloud-first products can already reach substantial distribution, although we found much less evidence of new Forge-native apps reaching the scale of the old Marketplace leaders.

Deep Clone for Jira is a useful example. It is cloud-only and has reached roughly 12,000 installs by solving a very specific administrative problem: duplicating Jira issues, epics, projects and configurations at scale.

That is already more installs than several famous legacy Marketplace products. codefortynine, the vendor behind Deep Clone, has also reached Atlassian's Platinum partner tier and expanded into related administrative products.

Other younger vendors show the same possibility at smaller scales. SaaSJet's Time in Status has more than 5,000 installs. Exalate has around 4,000 Marketplace installs and says more than 2,500 companies use its cross-system synchronization technology. JXL has won large customers such as Canva and Rivian by rebuilding the Jira issue-management experience around spreadsheet-like interaction.

What we do not yet see is a new generation producing many publicly verifiable $20-million individual apps. The Marketplace's biggest current businesses still disproportionately come from products that accumulated customers for five, ten or fifteen years.

Are the newest Forge apps already making millions?

Some Forge businesses are clearly reaching meaningful revenue, but the evidence does not yet show a broad replacement of the old Marketplace leaders.

Atlassian's own revenue-share policy gives us an unusually useful benchmark. Since 2026, eligible Forge partners can keep 100% of qualified revenue until they reach $1 million of lifetime Forge revenue. Atlassian explicitly designed the incentive around helping Marketplace businesses get through that first $1 million.

That does not mean every Forge app reaches $1 million. It shows that Atlassian now considers seven-figure app revenue a normal enough milestone to structure its ecosystem economics around it.

The installed-base data still tells a different story at the top. ScriptRunner, draw.io, Xray, Tempo Timesheets, JMWE and other leaders were established long before Forge became the strategic platform.

Forge is changing the economics for the next generation rather than instantly changing who makes the most money today.

The opportunity is genuine. Publishing a Forge app now, though, does not give a developer the same distribution advantage that early Marketplace vendors enjoyed a decade ago. There are more customers today, but also more than 6,000 competing apps.

Is AI creating a new class of Atlassian app winners yet?

Not yet. AI is rapidly becoming a feature inside successful Atlassian apps, but we found much stronger evidence of existing categories monetizing AI than of standalone AI Marketplace apps becoming the new revenue leaders.

Xray now markets AI-assisted test generation and prioritization. Tempo is building Rovo agents around portfolio and Structure workflows. Deep Clone, reporting products and other Marketplace vendors are adding AI assistance around existing administrative jobs.

The logic is straightforward. Existing apps already own valuable workflow context. Xray knows a customer's tests. Tempo understands time, projects and capacity. Structure understands Jira hierarchies. An AI function attached to this data has a clearer business case than another generic chatbot.

Atlassian is simultaneously putting AI directly into Jira, Confluence and Rovo. That creates a dangerous position for apps whose only differentiation is access to a language model.

For now, the strongest AI opportunity appears to be using AI to deepen an established workflow rather than selling “AI for Jira” as the whole product. We would not put any standalone AI app in the same commercial tier as Timesheets, ScriptRunner, Xray, draw.io or Structure based on the evidence currently available.

Are Marketplace install counts actually useful for finding profitable Atlassian apps?

Install counts are useful for identifying commercial candidates, but they are not revenue rankings.

Consider three products. draw.io has more than 60,000 Confluence installs. Xray has roughly 25,000 installs. Structure has around 13,000. If installs mapped directly to revenue, draw.io should dwarf Structure.

Yet Structure's former owner disclosed approximately $23 million ARR years ago. Enterprise portfolio-management software can command much higher effective revenue per organization than a broadly distributed diagramming utility.

Install counts also mix company sizes. Atlassian's Marketplace billing commonly follows the seat count of the underlying Atlassian product. One 10,000-seat enterprise can therefore be worth dramatically more than dozens of tiny Jira sites.

We should consequently use installs as evidence of distribution, then combine them with category economics, customer type, disclosed financials and vendor scale.

That is why an app with 4,000 enterprise customers can sometimes be more interesting commercially than one with 20,000 installations.

Which Atlassian app categories are producing the strongest businesses?

The strongest Atlassian app businesses cluster around expensive operational problems rather than novelty features.

Test management has Xray and Zephyr. Time and resource management has Tempo. Portfolio management has Structure and BigPicture. Workflow administration has ScriptRunner, JMWE and JSU. Diagramming has draw.io. Documentation governance has Scroll and Comala.

Integrations are another attractive category because companies will pay to keep two business-critical systems synchronized. Exalate, Salesforce connectors and Azure DevOps integrations illustrate this, although their public install counts are lower than the giant workflow and testing apps.

The common denominator is organizational dependency. Once hundreds or thousands of employees depend on a Marketplace app to submit timesheets, manage releases, generate audits, plan portfolios or maintain Jira workflows, removing it becomes a business project rather than an app uninstall.

Category	Commercially important examples	Why customers keep paying
Time / resource management	Tempo Timesheets	Billing, costing, capacity
Portfolio management	Structure, BigPicture	Cross-team planning and governance
Test management	Xray, Zephyr	QA history, traceability, CI/CD
Jira administration	ScriptRunner, JMWE, JSU	Embedded workflows and automation
Diagramming	draw.io	Persistent organizational documentation
Documentation	Scroll, Comala	Publishing and governance workflows

Are Atlassian's 2026 Marketplace fees changing which apps are attractive businesses?

Yes. Atlassian's current fee structure increasingly rewards fully Forge-native software and makes old Connect architecture less economically attractive.

Eligible Forge apps can currently keep 100% of gross Marketplace revenue until the partner crosses $1 million of lifetime Forge revenue. Above that threshold, Atlassian's standard Forge share leaves the vendor roughly 83%–84%, depending on the applicable period.

Connect economics are worse. Atlassian raised its cut and is moving Connect publishers toward a vendor share of around 75%. Data Center apps also generally return about 75% to publishers.

For a mature app producing $10 million of Marketplace revenue, a difference of roughly eight percentage points between Forge and a 75% vendor share represents around $800,000 per year. Architecture is becoming a financial issue, not simply a technical one.

That favors companies capable of funding difficult migrations. draw.io has moved aggressively toward Forge, while Atlassian is pushing the entire ecosystem away from Connect.

The paradox is that many of today's largest apps were built on the old architecture. They have the revenue to migrate. A smaller legacy vendor can face the same engineering cost without the same financial cushion.

So which Atlassian apps are making real money now?

The answer is clear: Atlassian has a genuine multi-million-dollar app economy, and the apps making the most credible money today are concentrated in test management, time tracking, portfolio management, advanced Jira administration, diagramming and documentation.

Structure gives us the strongest historical app-level proof: approximately $23 million ARR and EBITDA margins above 20% even before Tempo bought it. Tempo Timesheets now sits inside a 30,000-plus-customer software platform. ScriptRunner has roughly 35,000 installs and belongs to a group generating £325 million annually, although that company figure includes much more than the app. Xray has about 25,000 installs in an enterprise-heavy testing category. draw.io exceeds 60,000 Confluence installs. Appfire has turned a portfolio containing JMWE, JSU, BigPicture and dozens of other apps into a company with well above $200 million ARR.

We would place Zephyr, Easy Agile, Deep Clone, Scroll, Comala and several integration/reporting vendors in the next commercial group: clearly meaningful businesses, but with less public evidence allowing us to quantify current app-level revenue precisely.

What has changed is the definition of success. A $10,000-per-month Atlassian app remains a good small business, but it is nowhere near the ceiling anymore. The established winners have demonstrated that a Jira or Confluence extension can become a $5 million, $20 million or even portfolio-scale $100-million-plus software operation.

The catch is that the biggest winners today mostly own difficult, deeply embedded enterprise workflows. The evidence does not point toward tiny AI widgets or cosmetic Jira utilities as the main money makers. It points toward apps that effectively become another business system inside Atlassian.

App / product family	Evidence of commercial scale	Our current reading
Structure by Tempo	~$23M ARR disclosed at 2021 acquisition; ~13k installs now	Proven eight-figure app business
Tempo Timesheets	~27k installs; Tempo has 30k+ customers	Very large established business
ScriptRunner for Jira	~35k installs	Almost certainly among the largest Jira apps
Xray Test Management	~25k installs	Strong enterprise-scale business
draw.io for Confluence	~62k installs	Exceptional paid distribution
Appfire: JMWE / JSU / BigPicture etc.	Appfire $200M+ ARR; major apps from ~11k to ~17k installs	Portfolio-scale winner
Zephyr Essential	~10k installs; long-established commercial testing product	Durable multi-million-dollar candidate
Easy Agile products	Independent estimates put vendor revenue in the millions	Meaningful independent app company
Deep Clone for Jira	~12k cloud installs	One of the strongest newer cloud-first examples
Scroll suite	Multiple paid apps with thousands of installs	Established Confluence documentation business