NV Access

Blind and low-vision people install NV Access's free, open-source NVDA screen reader to navigate Windows when commercial screen readers cost over a thousand dollars, and web developers adopt it as a testing tool to catch accessibility failures before launch.

ENABLE Model location

• Navigator-side Compensations → Use Assistive Technologies
• Builder-side Interventions → Test for Accessibility

What it is

NVDA (NonVisual Desktop Access) is a free, open-source screen reader for Microsoft Windows, developed by NV Access, a non-profit founded in 2007 by Michael Curran and James Teh, both blind Australian software developers who met as teenagers at a music camp for blind children in Mittagong, New South Wales.¹ Curran began building NVDA in April 2006 while studying at Monash University because the dominant commercial screen reader, JAWS, costs between $90 and $1,475 per year depending on the license, more than many of the computers it runs on.² "Why do blind people have to pay so much when many of them are unemployed?" Curran asked.¹

In 2008, a grant from the Mozilla Foundation allowed Teh to work full-time on NVDA, a turning point that accelerated development from a side project into a viable assistive technology.³ By 2024, NVDA had become the most commonly used screen reader worldwide: the WebAIM Screen Reader User Survey #10, with 1,539 respondents, found 65.6% of screen reader users reported using NVDA, surpassing JAWS at 60.5%.⁴ The software is used by over 250,000 people across 175 countries, translated into more than 55 languages by volunteers.¹

NVDA also serves a builder-side function: web developers use NVDA as a QA testing tool to verify that their sites work with screen readers before shipping. Harvard Digital Accessibility Services, WebAIM, and the American Foundation for the Blind all publish NVDA testing guides. Because NVDA interprets web markup strictly, without the heuristics JAWS uses to compensate for missing ARIA labels or broken heading structures, it exposes exactly the barriers a blind user would encounter.⁵

NV Access is funded through corporate grants (Mozilla, Microsoft, Google, Adobe), individual donations, and accessibility consulting services. The software itself remains free. This funding model makes NVDA available to anyone with a Windows computer, but it also makes the tool that addresses precarity itself precarious, dependent on the continued generosity of the corporations whose products it exists to compensate for.

Why it matters

Ted Henter, blinded in a 1978 automobile accident, founded Henter-Joyce Corporation in 1985 and shipped the first version of JAWS in 1988 with a $180,000 investment from Bill Joyce.⁶ JAWS for Windows 1.0 shipped in January 1995. By 2003, JAWS commanded approximately 85% of new screen reader sales, according to Chris Hofstader, who served as VP of Software Engineering at Freedom Scientific from 1998 to 2004.⁷ In April 2000, Henter-Joyce merged with Blazie Engineering and Arkenstone to form Freedom Scientific, concentrating the commercial screen reader market in a single company. In 2015, Vector Capital, a San Francisco private equity firm, acquired Freedom Scientific and Optelec and combined them into Vispero. In 2017, Vispero discontinued Window-Eyes, the second major commercial Windows screen reader, which it had acquired through GW Micro, leaving JAWS without a comparable commercial competitor.⁸ The Americans with Disabilities Act of 1990 and the Section 508 amendment of 1998 created institutional demand for AT, channeled primarily through vocational rehabilitation agencies that became Vispero's most reliable buyers. Private equity consolidated a captive market and charged institutions whose clients had no viable alternative. Curran began building NVDA in 2006 directly against that arrangement, asking why blind people had to pay so much for a tool when many of them were unemployed.¹

American vocational rehabilitation (VR) agencies, funded through the Rehabilitation Services Administration and administered by states, buy JAWS for their clients and frame the purchase as an employment outcome investment. Freedom Scientific's own blog makes the institutional pitch explicit: "A rehabilitation agency investing in JAWS for a client is an investment that will be returned through a blind person being able to contribute as an employed taxpayer."⁹ Clients trained on JAWS within the VR system may find themselves reliant on a $1,150-to-$2,316 perpetual-license tool once they exit the program. The WebAIM Survey #10 shows what this institutional arrangement produces globally: JAWS leads only where state AT funding concentrates, in North America (55.5% primary usage vs. NVDA's 24.0%) and Australia (45.8% vs. 37.5%). NVDA leads in Europe (37.2%), Asia (70.8%), and Africa and the Middle East (69.9%).⁴ The American Foundation for the Blind, reporting 2024 Bureau of Labor Statistics data, found a 10% unemployment rate for people with visual impairment against 4% for people without disability, and a labor force participation rate of 44% against 75%.¹⁰ AT specialist certification programs and vendor support contracts reinforce JAWS's dominance through institutional inertia: agencies purchase what their staff know how to configure, what vendors certify, and what training programs teach.

Freedom Scientific, as Hofstader documented from inside the company, designed its patent strategy to suppress competition: then-CEO Lee Hamilton acknowledged targeting competitors with patents to "drop boulders in the roadmap of our competitors," and Hofstader reports that every major Freedom Scientific patent was ultimately found invalid, but only after competitors spent heavily defending against them.⁷ NVDA removes the $1,150-to-$2,316 per-seat cost that makes screen reader access contingent on institutional funding or personal wealth. The National Federation of the Blind's Access Technology Affordability Act, introduced as H.R. 3702 in the 118th Congress (2023-2024), proposed a refundable $2,000 tax credit for blind individuals purchasing qualified AT, citing 69.1% unemployment or underemployment among blind Americans and leading screen reader costs of $900 or more.¹¹ The bill did not pass. NVDA shifts some of this adaptation tax off blind users, but the development labor that keeps NVDA current runs on corporate grants, donations, and volunteer work, with no statutory underwriting. NV Access does not publicly disclose its annual budget, and its continued operation depends on the continued generosity of Mozilla, Microsoft, Google, Adobe, and individual donors.

The National Federation of the Blind's Access Technology Affordability Act advocacy record documents the community's reading of the cost barrier directly: the bill's stated rationale names 69.1% unemployment or underemployment among blind Americans and traces that figure to AT cost.¹¹ The NFB also maintains a dedicated "Low-Cost Screen Readers" resource page that promotes NVDA and other free alternatives, publicly treating the commercial market's pricing as a barrier the community must route around.¹² The blind community has built a dense infrastructure around NVDA's open-source core: 700+ community-developed add-ons in the NV Access add-on store, volunteer translation into more than 55 languages, and mailing lists and community podcasts that collectively function as an unpaid support channel for a tool with no paid help desk.¹³ A 2010 ACM SIGACCESS paper, "Disability studies as a source of critical inquiry for the field of assistive technology," documented that disability studies scholars are skeptical of technologies that "fix impairments" rather than challenging inaccessible environments, and that the AT field's primary orientation is toward individual adaptation rather than structural change.¹⁴ Community give-feedback to NV Access through bug reports, add-on development, and translation labor extends the tool's reach in ways that no commercial arrangement would fund, and it operates simultaneously as compensation and as pressure on the environments that require compensation in the first place.

Inaccessible digital infrastructure reaches into blind people's bodies through mechanisms that the COVID-19 pandemic brought into acute focus. Inaccessible QR-code authentication systems excluded blind users from restaurants, transit systems, and public spaces during the pandemic, and inaccessible online grocery platforms pushed them toward reliance on human assistance for tasks that digital infrastructure was supposed to make independent.¹⁵ A 2021 study in PMC (PMCID 8423669) found that people with disabilities showed smaller internet adoption increases than people without disabilities during the pandemic, with blind-specific barriers documented across health information portals, grocery sites, and government services, and identified digital inaccessibility as structural technoableism, in which the expectation that users adapt to inaccessible design concentrates cognitive and psychological burden in disabled populations.¹⁵ The WebAIM Survey #10 found only 37.6% of screen reader users with disabilities employed full-time against 83.2% of non-disabled respondents.⁴ The mechanism linking AT cost to embodied harm is traceable: inadequate AT access reduces employment access, employment exclusion produces precarity, and economic precarity concentrates the chronic health consequences of poverty, social isolation, and deferred care in populations the built environment already places at elevated risk. The WHO estimates annual global productivity loss from unaddressed vision impairment at approximately $411 billion (PPP), with approximately 90% of the visually impaired population living in settings where institutional AT funding does not reach.¹⁶

NVDA reached global parity with JAWS in the WebAIM Survey #10: 65.6% of screen reader users report using NVDA against JAWS's 60.5%, a reversal achieved by free software maintained through charity against a product backed by private equity.⁴ NV Access's published 2025-2026 roadmap includes on-device AI image description that runs locally rather than transmitting screen content to cloud servers, machine learning-based UI element recognition for inaccessible interfaces, and a 64-bit runtime migration.¹⁷ NVDA's capability now exceeds the semantic quality of the web it navigates. The WebAIM Million's 2024 analysis of the top one million homepages found WCAG 2 failures on 95.9% of pages, with an average of 31.0 distinct accessibility errors per page.¹⁸ On-device AI can describe an image that builders left unlabeled. It cannot reconstruct the heading structure a developer chose to omit, the button label a designer skipped, or the keyboard trap a front-end engineer never tested before launch.

warning

Sighted developers who run NVDA can verify that headings exist, that buttons are labeled, and that focus order is logical, but they cannot replicate the cognitive model of someone who navigates by sound alone. Screen reader users develop spatial memory, listening shortcuts, and navigation heuristics that no testing checklist captures. Developer testing with NVDA catches markup failures, but it does not catch experience failures. Organizations that treat developer screen reader testing as a substitute for testing with blind users are replacing one form of absent care with another.

Real-world examples

In the news

Blind inventors revolutionize computer access (October 2010)
-- Phys.org

• Michael Curran and James Teh won the "Les is More" award on ABC's New Inventors television show for NVDA. At the time, NVDA had over 50,000 downloads in 27 languages. The article noted that commercial screen readers cost more than $1,000, a barrier that NVDA eliminated entirely. The award recognized what a Mozilla Foundation grant had made possible: a full-time development effort that turned a student project into the world's leading screen reader.

Adobe and NVDA Partnership Improves Accessibility (2012)
-- Knowbility

• Adobe entered a formal partnership with NV Access to improve NVDA's compatibility with PDF accessibility features and the PDF/UA standard (ISO 14289). Adobe issued a "Statement of Support for Open Source Assistive Technology." The partnership illustrates builder-side care: a corporation investing in the assistive technology its users depend on rather than leaving compatibility gaps for blind users to navigate alone.

WebAIM Screen Reader User Survey #10 (January 2024)
-- WebAIM

• The definitive independent survey on screen reader usage found NVDA had become the most commonly used screen reader worldwide (65.6%), surpassing JAWS (60.5%). Respondents rated CAPTCHAs as the most problematic accessibility barrier, a failure of builder-side design that no amount of screen reader sophistication can compensate for. 97.6% of NVDA users reported being somewhat or very satisfied with the tool.

• In 2008, the Mozilla Foundation awarded NV Access an $80,000 grant that allowed James Teh to work full-time on NVDA, a builder-side investment in assistive technology infrastructure that transformed the project's trajectory.³
• In 2012, NV Access received the American Foundation for the Blind's Access Award, recognizing NVDA's impact on digital accessibility for blind users worldwide.¹⁹
• Microsoft published a feature profile of NVDA on its Microsoft Unlocked platform, reporting over 250,000 users in 175+ countries and acknowledging NVDA as critical infrastructure for blind digital participation.¹
• Fernando, a blind job seeker in Brazil, wrote to NV Access: "Companies never hire us because they do not want to bear the cost of the screen reader. Making use of NVDA allows me to enter the labor market without requiring the company to bear the costs acquisition of a paid screen reader."²⁰ Unlike JAWS, which is tied to a licensed machine, NVDA can run from a USB drive on any Windows computer. Teh described it as a tool that "can be copied to a USB stick, which can be used on any PC at school or university, with no installation required."³
• Bhavya, a blind high school student in India, wrote: "NVDA has transformed my life. Although I am still a high school student, I have been able to win Model United Nation conferences, achieve distinctions in mathematics and science related competitive examinations... largely due to the powers and abilities of NVDA, at no cost."²⁰

What care sounds like (builder-side interventions)

Care at the design and development level involves building software that screen readers can fully navigate:

• "Why do blind people have to pay so much when many of them are unemployed?" -- Michael Curran, explaining why he built NVDA as free software¹
• "Accessibility and equitable access is a right and should be available to everyone." -- NV Access²¹
• "We test every release with NVDA and JAWS before shipping."
• "We're partnering with NV Access to ensure our PDFs are readable by screen readers." -- the premise of Adobe's 2012 partnership¹⁹
• "We wrote heading structures and ARIA labels so screen reader users can navigate by landmarks, not by guessing."

What neglect sounds like (builder-side interventions)

Neglect involves assuming that accessibility is someone else's problem, or that screen readers will compensate for absent care:

• "Screen readers will figure it out, right?" (They won't. NVDA reads what builders write. If the markup is empty, so is the experience.)
• "It looks fine, so we assumed it works." (Visual appearance tells you nothing about screen reader compatibility.)
• "We don't have blind users." (The WebAIM survey found 250,000+ NVDA users worldwide. You have blind users. You just can't see them leaving.)
• "JAWS handles our site fine." (JAWS uses heuristics to guess at bad markup. NVDA does not. Testing only with JAWS can mask failures that blind NVDA users hit every day.)⁵
• "We'll add ARIA labels in the next sprint." (Meanwhile, a blind person is tabbing through 47 unlabeled buttons.)

What compensation sounds like (navigator-side compensations)

Compensation describes the labor blind people undertake when builders ship inaccessible software and NVDA is the only way through:

• "I use a screen reader, but this site has unlabeled buttons everywhere. I have to guess what each one does."
• "I had to install a screen reader just to apply for this job. The application portal has no heading structure, so I Tab through every element on the page."
• "I carry NVDA on a flash drive because I never know if the computer I sit down at will have a screen reader installed."
• "I never know which website will be usable today. Some days I can buy groceries online; some days the checkout breaks and I have to call someone."
• "The screen reader works. The website doesn't. But everyone assumes the problem is my screen reader."

---

All observations occur within the context of the global assistive technology ecosystem, where NV Access's free screen reader has become the primary tool through which blind people in most of the world access Windows computers and the primary tool through which many developers discover whether their software is accessible at all.

Footnotes

1. Microsoft Unlocked: An open-source screen reader for all ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

2. UXPin: NVDA vs. JAWS: Screen Reader Testing Comparison ↩

3. Phys.org: Blind inventors revolutionize computer access ↩ ↩² ↩³

4. WebAIM: Screen Reader User Survey #10 ↩ ↩² ↩³ ↩⁴

5. CSS-Tricks: A Brief Introduction to JAWS, NVDA, and VoiceOver ↩ ↩²

6. Ted Henter Oral History, Part 1 - American Foundation for the Blind ↩

7. The Death of Screen Reader Innovation - Chris Hofstader ↩ ↩²

8. Window-Eyes Discontinued - Cool Blind Tech ↩

9. Extending Our Reach: Making JAWS More Available - Freedom Scientific Blog ↩

10. Employment Statistics for People Who Are Blind or Have Low Vision - American Foundation for the Blind ↩

11. Access Technology Affordability Act - National Federation of the Blind ↩ ↩²

12. Low-Cost Screen Readers - National Federation of the Blind ↩

13. NVDA Add-on Store - NV Access ↩

14. Disability studies as a source of critical inquiry for the field of assistive technology - ACM ASSETS 2010 ↩

15. Effect of Digital Divide on People with Disabilities During COVID-19 - PMC PMCID 8423669 ↩ ↩²

16. Blindness and Visual Impairment - World Health Organization ↩

17. NVDA Roadmap - NV Access ↩

18. WebAIM Million 2024 - The accessibility of the top 1,000,000 home pages ↩

19. Knowbility: Adobe and NVDA Partnership Improves Accessibility ↩ ↩²

20. NV Access: About NVDA ↩ ↩²

21. Digitally Connected: Michael Curran and James Teh ↩