diagrams.net teardown: 10.62M visits a month, free with no paywall

diagrams.net, the free diagramming tool also known as draw.io, draws an estimated 10.62M visits a month (Semrush, October 2026). That makes it the largest diagram tool in our database and the fifteenth largest site we track. It charges nothing on its own site. There is no paid tier, no account wall and no Enterprise plan for single sign-on. The homepage says it in two words: "No Paywalls". The company behind it, draw.io Ltd in England and draw.io AG in Switzerland, says the paid product lives somewhere else entirely: in the Atlassian ecosystem, where Confluence and Jira customers pay for the same diagramming inside tools they already pay for. The question this teardown answers is the obvious one. How does a site with 10.62M monthly visits and no price tag pay for a professional team?

What problem does diagrams.net solve, and for whom?

A flowchart, an architecture diagram or a UML sketch is a document that needs to exist and needs to be editable by the next person who opens it. The usual answer is a seat in a paid diagramming tool, or a file format that only the paid tool can open. diagrams.net's answer is a page that opens in the browser, draws the diagram, and saves a file anyone can open. The homepage we read on 16 November 2026 opens with "Security-first diagramming for teams" and closes the first screen with "No account required. No credit card. Just open and create."

The audience is anyone who draws a diagram at work, which is a wide net, and the traffic figure shows what that net catches: 10.62M estimated visits in October 2026, the largest diagram tool we track and more than twice the next largest diagramming site in the database, Excalidraw at 4.13M to 4.57M. The country split is worth noting because it is flatter than the PDF tools: the United States leads at 22.08%, but Indonesia at 9.52% and Vietnam at 6.66% follow, and no single market carries the site.

The product also serves a second audience that never visits diagrams.net at all: developers who embed diagrams in docs, wikis and applications. The homepage shows an iframe embed snippet for an editable diagram viewer, and the about page describes draw.io as a technology stack for building diagramming applications, not only as a website. That second surface matters for the business model, because it is where the paid version lives.

The product and its first screen

The homepage we read on 16 November 2026 is built around six promises, each given its own block: backwards compatibility, sharing with everyone, real-time collaboration, privacy first, no paywalls, and independence. Diagram files created in 2005 load in today's app, the site says, so work is never obsolete. Anyone can open and edit draw.io files on any platform, with no licence worries. Real-time collaboration comes with shared cursors. Privacy is stated as a hard rule: store your data wherever you want, because the company cannot access it, ever.

The "No Paywalls" block is the one doing the most work. It reads, in full: "Apache 2.0 licensed source code. No artificial scarcity, no Enterprise tier for SSO." The "Independent" block adds the funding position: "Not VC-funded. Private company focused on great software, not inflating valuations." Together the two blocks are the pitch. Most tool sites use the homepage to justify a price. This one uses it to justify the absence of one.

What the homepage does not do is show the tool. There is no live editor on the first screen, no template gallery and no customer logos. The page sells the terms of the deal, not the drawing experience, and the terms are the product for a visitor who has been burned by a paywall before. The one piece of product on the page is the embed snippet for developers, which points at the audience that actually pays: teams whose diagrams live inside Confluence and Jira.

Pricing: there is no price, and that is the business model

diagrams.net has no pricing page and no paid tier. The homepage says the whole product is free, the source code is Apache 2.0 licensed, and there is no Enterprise tier for single sign-on. There is nothing to upgrade to on diagrams.net itself, and no trial to expire, because there is nothing to try.

The money, according to the company's own about page, comes from one place: the Atlassian ecosystem. The about page says the model is "based on the viral effect of a free application", and that the company picked "one commercial ecosystem to charge for the application where the host product is pay-for and the profit margins greatest, and that was the Atlassian ecosystem." Confluence and Jira customers pay for draw.io apps inside tools they already pay for, while the web version stays free. The about page adds the rule that governs the split: "When companies pay us money it should be because we add value, not because they are locked in."

The about page also states the mission in one line: "provide free, high quality diagramming software for everyone". The same page gives the reasoning behind the model: diagramming software is a billion-dollar-a-year market, in the company's own words, but "nothing is happening in the field that justifies this sales volume", and SaaS vendors, in its view, charge monthly for access to data they also store. That is a position, not a neutral fact, and it is quoted here as the company's own statement. It is also the whole pricing strategy: be the free version good enough that the paid version is bought for convenience inside Atlassian, not for access to the tool.

How does diagrams.net get its users?

Not publicly disclosed. diagrams.net publishes no acquisition mix, and our database holds no verified channel data for it, so this section is short on purpose. Anyone who tells you its traffic is exactly X percent organic search is guessing, however plausible the guess sounds.

Three things are observable without inventing numbers. First, the product shape gives the site a natural distribution channel that costs it nothing: every diagram a user shares is a file the next person can open and edit for free, in any browser, with no licence. The about page says this directly: "Our model is based on the viral effect of a free application. When you share a draw.io diagram, you know anyone can open and edit it." A shared file is an invitation that does not ask the recipient to pay or sign up. Second, the embeddable viewer puts editable diagrams inside docs, wikis and applications, which is a second surface that reaches people who never set out to visit a diagramming site. Third, the traffic is large enough to be structural: 10.62M visits in October 2026 per Semrush, the fifteenth largest figure in our database. A site does not reach that level on one campaign, and a free tool with no account wall is built to be linked to, which is the cheapest acquisition channel there is.

The one channel we cannot see is search, and we will not guess at it. What we can say is that the site's own words point at sharing, not advertising, as the engine, and that a free tool with an open file format is built for exactly that.

From visit to payment: there is no funnel on this site

There is no conversion funnel to dissect on diagrams.net, because there is nothing to convert to. A visitor opens the site, draws a diagram, and leaves with a file. No account, no credit card, no trial clock. The conversion that matters happens elsewhere, in the Atlassian marketplace, where a Confluence or Jira administrator pays for the same diagramming inside a tool the team already uses. We cannot see that funnel from outside: no marketplace revenue, no install count and no conversion rate is published in the sources we checked.

The honest way to describe the funnel is as a two-step bet. Step one is the free web app, which has to be good enough that millions of people use it and share files made in it. Step two is the paid Atlassian app, which has to be convenient enough that a team already paying for Confluence pays a little more for diagrams inside it. The free site is the top of the funnel, and the paid product is sold to a different buyer in a different place. A founder copying this model should know which part they are copying: the free tool alone, without the paid ecosystem to charge in, is a hobby, not a business.

Retention: what keeps people coming back?

Not publicly disclosed. diagrams.net publishes no churn, renewal or cohort figures in the sources we checked. The structural argument is still worth making, because the site makes it itself. A diagram is a document that gets reopened, edited and passed on, and a file format that has loaded since 2005 is a reason to return to the same tool rather than learn a new one. The homepage's backwards-compatibility promise, "Diagram files created in 2005 load in today's app. Your work is never obsolete", is a retention argument written as a feature.

The second retention mechanism is the absence of lock-in, stated as a promise on the about page: "We promise you we won't hide your data from you and will always give you some way to open and edit that data, at no cost." A user who knows they can leave is a user who does not have to think about leaving, and a file anyone can open is a file that keeps circulating. Whether that plays out as repeat visits to diagrams.net or as diagram files opened inside Confluence, we cannot tell from public data. Both would fit a free tool whose paid product lives in the Atlassian ecosystem. Neither is proven by the traffic figure.

Why diagrams.net is still standing

Our read, stated as a position you can argue with: diagrams.net survives because it refused the trade every other tool in our database made. The others charge for the tool and use a free tier to fill the funnel. diagrams.net gives the tool away, charges for the same tool inside a paid ecosystem, and lets the free version do the marketing. That only works if the free version is genuinely free, and the site's homepage is a list of the ways it proves that: open source, no paywall, no account, no Enterprise tier, your data stays yours, files from 2005 still open. Each promise removes a reason to distrust a free tool, and distrust is the main reason a free tool fails.

The traffic figure backs the position, with its label attached: an estimated 10.62M visits in a single month (Semrush, October 2026) is not a side project hanging on. It is the largest diagram tool we track, more than twice the next one, Excalidraw at 4.13M to 4.57M, and the fifteenth largest site in a database of forty-five. A free product with that much usage and a paid product in exactly one ecosystem is a defensible place to stand, as long as the ecosystem keeps paying.

What you can copy, and what you cannot

Copyable this week

• Make the free version the whole product, not a sample of it. diagrams.net's homepage lists what is not charged for, and the list is the pitch. A free tier that is obviously complete is a free tier people trust.
• Let the file be the invitation. A diagram anyone can open and edit for free is distribution that costs nothing, and it works because the format is open and the viewer is free.
• Promise backwards compatibility in plain words. "Files created in 2005 load in today's app" is a sentence a user can test, and a promise like that is worth more than a feature list.
• Charge where the host product is already paid for. If your tool lives inside a platform people pay for, that is where the paid version belongs, not on your own site.
• Say who you are not. "Not VC-funded. Private company focused on great software, not inflating valuations" is a position that costs nothing to print and tells a buyer exactly what to expect.

Not copyable

• The Atlassian ecosystem as a revenue source. It took years of a free product spreading through shared files to be worth paying for inside Confluence and Jira. You can build a free tool this week. You cannot start with that distribution.
• The search presence behind an estimated 10.62M monthly visits. That compounds over years of a free tool being linked to. You can start the same flywheel; you cannot start where diagrams.net is on it.
• The trust position. Open source from the start, no paywall, no lock-in: that is a record, not a landing page, and a new site claiming it has not earned it yet.

Risks and gaps in this teardown

• The traffic figure comes from a single source for the month quoted, Semrush, and is a third-party estimate. diagrams.net publishes no official MAU or DAU.
• The about page quoted here (diagrams.net/about) was last crawled on 27 April 2026, and the operator details, mission statement and business model description come from that page. The homepage statements were read on 16 November 2026.
• The operator is draw.io Ltd (England, company no. 04051179) and draw.io AG (Switzerland), per the about page. An earlier third-party report named JGraph Ltd as the operator; the about page is the source used here because it is the company's own statement.
• The launch year of diagrams.net is not stated in the sources we checked. The homepage's reference to diagram files created in 2005 is a claim about file compatibility, not a founding date, and no launch year is claimed here.
• The claim that draw.io is "the world's most widely used browser-based end-user diagramming software" is the company's own statement from its about page, quoted as such, not independently verified.
• The Atlassian marketplace revenue, install counts and conversion rates are not publicly disclosed in the sources we checked. The business model description is the company's own account of how it charges, not a verified revenue figure.
• Acquisition mix, conversion rates and retention are not publicly disclosed. Sections above say so rather than estimating them.
• Country breakdown comes from Semrush for October 2026 and covers the top three countries only. The remaining share is not broken out in the sources we checked.

Similar sites in the database

Site	Category	Monthly visits	Source
Excalidraw	Design tools	4.13M to 4.57M	Semrush and Hypestat, Oct 2026
Whimsical	Design tools	1.27M	Semrush, Oct 2026
tldraw	Design tools	1.05M	Semrush, Oct 2026

Compare diagrams.net with Excalidraw and Whimsical side by side. The full database, with working filters, is on the teardowns index.