Open Tax Technology Alliance launches with free 1040 engine

Tax solution specialists Filed and Crimson Tree Software have teamed up to launch the Open Tax Technology Alliance, a new organization dedicated to promoting and supporting open source software in the accounting technology space. As one of the Alliance's first initiatives, it has released a free 1040 tax calculation engine anyone can use, as well as an open data standard for partnership returns that bypasses PDFs in favor of digitized information. 

Processing Content

The organization was born out of frustration with the current accounting software landscape, and the solutions sold to handle the various frictions in accounting workflows.

"People make money on friction and that causes inertia, it causes the industry to not want to change," said Tom O'Sullivan, co-founder and chief engineer at Crimson Tree Software. "But there's vast numbers of people who want these problems to go away. We want to make this type of problem go away and do it in a fashion that brings many people to the table instead of being to the benefit of just a few."

Tom and Leroy Open Tax Technology Alliance
Leroy Kerry and Tom O'Sullivan, founders of Open Tax Technology Alliance

Leroy Kerry, co-founder and CEO of Filed, said this has led to a closed loop system of certain software companies that have essentially treated the problems of the accounting profession as a long-term investment, which has caused frustration for practitioners. Firms want more flexibility and openness. He and O'Sullivan discussed this need.

"When we spoke initially, we shared this common theme that we were also frustrated by these problems, which comes not just from technology but a need to rethink things from the bottom up," said Kerry.

Broadly, their organization will engage in advocacy to encourage firms to adopt open source software and open data standards, educate people on both, and provide direct source code and implementation guidance for open source software solutions. The two conceded that they're still working out a lot of the specifics, but one thing they know for sure is the organization itself will be not-for-profit, and the open source solutions and standards it releases will never be monetized.

"There's room for a lot of people to earn a good living in an industry or market that is also more open," said O'Sullivan. "Our perspective is that a lot of us are getting older. We have been really successful, so we thought let's give something back that could potentially catalyze positive change."

Open Tax

One of the first things the organization is releasing is Open Tax, described as an automated Form 1040 calculation engine that ships as a single binary that runs on MacOS, Windows and Linux with no account or cloud dependency. It covers Form 1040 for tax year 2025 across 186 registered nodes and 131 input types, including W-2s, the 1099 series, K-1s and Schedules A, C and E. It also checks returns against the IRS MeF business rule set. While it does not actually e-file for you, it does export modernized e-File XML.

"All firms can actually now utilize a more secure system that can actually give them the ability to run any LLM of their choice locally, which means they don't even have to connect it to the cloud, or they don't have to connect it to the internet," said O'Sullivan. "They can run all of their calculations on the local machine with no prying eyes, no one asking for a payment."

The engine should not be thought of as a tax product unto itself but as something people can decide to build into other tax products, whether incorporating it into an existing solution or using it to make an entirely new platform. While doing this can make it usable to a human, it can also be set up to be used directly by AI agents to make 1040 calculations themselves, perhaps over a Model Context Protocol server connected to Claude or another frontier model or perhaps using one's own local or cloud-based models. Kerry stressed that the engine has parsed all IRS test cases and its code is open and completely auditable for security, reliability and accuracy.

O'Sullivan noted that it can also be used to compare results and catch possible mistakes. He has encountered instances where one tax engine disagrees with another and, after investigation, found this was due to a bug or glitch in the code. He believes an open source solution like this would let someone cover their bases more effectively.

"What is exciting to me about having an open tax engine is the idea that I can have my agentic software compare the results of a tax return prepared by my CPA versus a self-prepared tax return, isolate the differences, and read the source code to determine if there's a number that's in dispute and how that number was computed. That's something that's completely unavailable in the current regime," he said.

He added that it allows firms to not just enact a process but to own the process itself as well as the software executing it.

"If a firm decides 'we want to own our own tax software and actually control our own destiny,' which I am hearing a lot, which is also part of the initiative driving this, they can now do that. They'll use what they feel is 'costworthy,' auditable and, most importantly, secure," he said.

This also is what allows people to have confidence in the engine and trust its calculations. It is not a black box; people can see how it operates. This sort of transparency, he said, is rare in the accounting software world, but is something he felt was important to deliver to the profession.

"You have these intricate calculations that have to take place sometimes, but generally speaking it is a stovepipe: Data goes in one end, other data comes out the other end. You have no idea what happened in that process. It's a black box. Our basic premise has been complete transparency. Any calculation that's done, you can pick out. … That commitment to transparency is what allows people to build confidence on the results that are coming out of your systems, and so what we are trying to do is not force, but set an example for transparency that should be demanded of our industry," he said.

The only rule is that the license is open, so if someone builds on these properties, they must be open source as well. Not only is the engine open source, but anything developed using the engine must also be open source.

"You can imagine someone pulling something, improving it," said Kerry. "You want to give it to someone else other than just yourself? You must push it back in. That's the rules. That's the licensing. That's just how open source code works. It's a system that governs itself, and it's a community of people that go and build upon it, that govern its effectiveness by building and making it better."

Kerry added they have other forms on the roadmap, such as 990s, 1041s, 1065s and 1120s. But he added that hobbyists have already found the engine and are already working on their own enhancements, which he said is ideally what you want with a project like this.

"So long may that continue," he said.

Open Tax Document standard

The alliance also released the Open Tax Document data standard, made for digital partnership returns. The data standard, freely available for anyone to use, was made specifically to address the extremely complex and cumbersome nature of partnership tax workflows, which often require a lot of PDFs to be passed around, creating not only friction but the possibility of error with each handoff. The data standard is meant to bypass this problem by digitally encoding the information versus trying to extract it from the PDF. O'Sullivan said it was ironic that people have created complex systems for partnership computations and yet people still structure their outputs in the form of what is still essentially a human-readable print format versus a data interchange format.

"Every step in that chain should be exchanging data, but instead we're exchanging paper," he said.

Encoding the information not in PDF but in its own digital data interchange format solves the problem at the root. The standard represents not only the K-1 and K-3 itself but all the attached footnotes, creating a comprehensive package of structured data in machine-readable format with what he said was "its own ontology, its own taxonomy."

"We think there's an opportunity to dramatically simplify a certain giant time-consuming problem that affects the industry," he said. 

Putting the data in a machine-readable data format also enables easier AI integration as agents can now directly access the data versus going through steps to extract it. This allows people to skip extremely complex steps in the workflow that have long troubled firms. 

Right now, tools from Filed and Crimson Tree Software are compatible with the data standard, but they intend for it to go much further than that. Kerry noted that both Filed and Crimson Tree work with many accounting firms, ranging from local practices to some Big Four firms. Ideally he and O'Sullivan would like to see at least one of the Big Four adopt the OTD standard and at least one in the next five largest. They'd also like to see large organizations like hedge funds and private equity funds, as well as family offices, adopt it as well. Not only is the standard specifically applicable to them, but their size would lend major visibility and inertia to even wider adoption.

"You have these free form footnotes where every investment fund has their own slightly different format. … What's missing there is an ontology they can map, like private equity firm one's footnotes to private equity equity firm two's footnotes, and so on. And so, what we would invite is exploration and discussion and a real thought process around what an ontology and taxonomy for these footnotes should actually look like," he said. 

While it is a standard, they won't seek to govern it like one might for accounting or audit standards, as it will be more aligned with the open source culture, so it is expected that many people throughout the world will be contributing to it. However, supported implementations must preserve information and reproduce equivalent normalized YAML (a human-readable data serialization language).

Free even at a cost

Both Crimson Tree and Filed are companies that produce tax software, so one might assume they are concerned that the free products they develop and support will compete with their own offerings. However, while Kerry and O'Sullivan concede that it might indeed create more competition, this misses the point. This is not some play to expand market share or find some new monetization strategy but rather to create things the entire profession can use. Both, at different times, described it as a "gift" to the industry, something that everyone can benefit from, including themselves and their competitors.

"The industry is ripe for a reformation," said O'Sullivan. "I think that many industries are facing this kind of identity crisis in the modern era, and the progress that's been made over the last couple of years broadly is less than what it should have been. I think that what's required is some gifts like this to catalyze change, to not force people but to make people think maybe there is a better way to do this,"

Kerry was blunt in noting that if they were thinking selfishly they wouldn't be doing this project at all. But he also noted that technology itself is only a small piece of what their companies do, so releasing free software doesn't really threaten their businesses, which are built on much more than just their tools.

"If i go into a room and tell people, 'Hey, I'm giving you my technology for free,' sure my investors may also have questions, but this isn't everything that we do," said Kerry. "It's one sliver of [our] companies. … Our gift is, whatever may happen in the future, we've at least given something to the profession, letting us make it better and establish precedents going forward."

Part of the precedent revolves around transparency and verifiability. Kerry lamented there are a lot of unproven claims regarding the accuracy and reliability of certain solutions that, on their face, might seem ludicrous. Open source solutions like the ones they are promoting, however, allow people to actually test the claims for themselves versus taking a vendor's word for it, something he'd like to see more of over time.

"Once we let you under the hood and test it yourself and see what the outcomes are, for me that is everything. This actually goes to the benefit of companies, so we hope everyone then starts to be open and more honest. It's one thing to make claims, it's another thing to have to prove it. The moment you open things up, you have proof, and therefore your claims can be validated by those who test it," he said. 

Another is for firms to remember where the real value lies. O'Sullivan said the calculations are only a small part of an actual tax engagement. When a client hires a tax firm, they're not just paying for the calculations, they're paying for a complete service that also includes judgment, insights and structure. As automation continues to improve, he said, the value of the pure compliance component is trending toward zero. While releasing a free tax engine and open data standard will contribute to this trend, and perhaps might undermine some incumbent players (even themselves, he conceded), he is unconcerned as it will ultimately drive higher value activities.

"Broadly, it should at first negatively impact the economics. It's deflationary with respect to service fees and license fees for software. But over the long term, we believe it is the opposite, that it will drive value in the work that the clients value, not in the work that only the vendors value," he said.

Overall, Kerry stressed that the purpose of the organization, tax engine, data standard and the entire project is ultimately to provide resources that can improve the profession as a whole for the benefit of its individual practitioners.

"We're missionaries, not mercenaries," he said.


For reprint and licensing requests for this article, click here.
Technology Practice management Artificial Intelligence Tax planning Partnerships
MORE FROM ACCOUNTING TODAY
Load More