Fixing Michigan’s Fragmented E‑Filing

Editor’s note: The author is an editor of the Journal. The views expressed, and the MUEFS project described, are the author’s own.

It was a Tuesday afternoon when the request arrived. A partner needed a subpoena domesticated in a Michigan county that still had no e-filing system. The clerk offered two choices: drive there or mail it. The courthouse was three hours away.

Six hours on the road, plus parking and waiting. Or risk the mail with something time-sensitive. The rest of Michigan kept moving. One lawyer cannot practice across an entire state from behind a desk.

Michigan has eighty-three counties. Its size turns ordinary filings into logistical problems.

MiFILE has brought real progress. It now operates in 127 trial courts.¹ Yet attorneys still encounter different interfaces in different counties, filing types that remain unavailable online, and entire courts reachable only by car or envelope.

The system that was meant to solve fragmentation has itself become fragmented.

That is why I built something else.

The result is MUEFS, the Michigan Unified E-Filing System.² It is a working proof of concept — not a toy or a wireframe, but a fully navigable application where every described feature actually runs. It was built using modern, accessible development tools — and that fact is itself part of the argument. It is a demonstration of what a single, modern platform could deliver across all eighty-three counties and every level of court. It is not a finished product. It is not a bid for a state contract. It is an answer to a simple question: can Michigan do better than the current patchwork?

The short answer is yes.

The problem with patchwork

Credit where it is due. MiFILE was a significant step. Before MiFILE, e-filing in Michigan barely existed. SCAO and its vendor partners brought dozens of courts online and gave attorneys a way to file without a trip to the clerk’s window. That matters.

But the system grew the way government technology often grows: in pieces, under contract constraints, on timelines set by budget cycles rather than user needs. The result is a platform that works court by court, not state by state. An attorney filing in Wayne County and then in Washtenaw encounters different interfaces, different quirks, different gaps. Some filing types in some counties still are not online at all.

The upload limit is 25 megabytes, which sounds like a lot until you scan an exhibit binder or a large discovery production. The system does not prompt you when a motion needs a brief or a proposed order. It does not warn you about companion document rules. These are not fatal flaws. They are the friction of a system built to get courts online, not to make filing easy.

What a unified system looks like

MUEFS takes a different approach. One login, one interface, every court in the state. The platform covers Circuit, District, and Probate Courts, Court of Claims, four Court of Appeals districts, and the Supreme Court. An attorney picks the court, picks the case, and files. The system handles routing, validation, fee calculation, and service.

Under the hood, the architecture uses an adapter layer designed to connect to existing case management systems like JIS and Tyler Odyssey. Those adapters are not yet wired to live systems — this is a proof-of-concept, not a production deployment. But the pattern means no court would need to rip out its existing software.

The underlying stack is worth mentioning. MUEFS is built on FastAPI, React, Docker, Keycloak for authentication, and MinIO for document storage — a modern, maintainable combination that any competent development team could pick up and extend. This is not an incidental detail. It matters because it shows the project is built on a foundation that could realistically scale, not just a demonstration that works once under ideal conditions.

The platform checks MCR 1.109 and SCAO formatting requirements on its own.³ Upload a document, and the system confirms it is a text-searchable PDF, flags PII, and checks the format. These are things attorneys now do from memory or from habit. A good system should do them by default.

Features that earn their keep

Some specifics. MUEFS includes a catalog of 135 document types drawn from the Michigan Court Rules. While the catalog itself is not court-specific, the system applies court- and case-type-specific filing requirements so each court enforces its own required companion documents in context. File a motion, and the system prompts for the expected companion documents — a brief in support under MCR 2.119(A)(2) and a proposed order (standard practice under MCR 2.602).⁴ File a discovery motion, and it prompts for a meet-and-confer certification under MCR 2.313.⁵ These are companion document rules that every litigator knows — and that every litigator has, at some point, missed on a rushed Friday afternoon.

The filing wizard walks self-represented litigants through the process in plain language. It does not assume they know what a “case type” means or which court to pick. It is built for the tenant answering an eviction notice or the grandmother seeking guardianship — people who meet the court system without a lawyer and cannot afford to get a step wrong. For attorneys, the interface is faster: open a case, click “File with Court,” and the system pre-fills the court, case title, and case type, so you only need to select the document and attach files. Case favorites let you bookmark the matters you touch every day. Drafts save on their own. A browser crash costs you nothing.

Fee waivers under MCR 2.002 are built into the payment flow.⁶ A filer who qualifies need not file a separate motion or track down a different form. The option appears at the right point in the process. Small thing. But for someone who cannot afford filing fees, it is the difference between access and a closed door.

On the clerk side, the review queue supports batch acceptance, tracks how long filings have waited, and provides common rejection reasons as one-click options. Fifty filings a day is normal. Clerks should not need to type the same rejection reason fifty times.

The upload limit is 100 megabytes. Four times what MiFILE allows. Not because bigger is always better, but because a litigator should not need to split a deposition transcript into four uploads to clear an arbitrary cap.

Case records are also searchable without an account. Anyone can look up a case by number or party name, view the docket, and download public documents. Sealed and confidential filings stay restricted to the filer and court staff. This is how public records should work: open by default, protected where it counts.

What courts and SCAO stand to gain

Most of the features above are pitched at attorneys. But the strongest case for a unified system may be the one aimed at the people who run the courts.

Right now, SCAO manages a patchwork of vendor contracts, each covering a subset of courts with its own terms, its own pricing, and its own update cycle. A single platform replaces that with one system to maintain, one set of training materials, and one line item in the budget. Courts that are not yet on MiFILE could come online without a separate procurement process. Courts already on MiFILE would not lose ground; the adapter layer is built to work with existing case management systems, not replace them.

The data argument matters too. Today, there is no easy way to pull statewide filing data across all courts. How many filings does the 3rd Circuit handle per month versus the 6th? What is the average time from submission to acceptance in Ingham County? Which courts have the highest rejection rates, and why? A unified system captures all of that in one place. SCAO would have real numbers to inform policy, allocate resources, and report to the legislature.

Pre-submission validation also helps clerks and courts directly. When the system catches a non-searchable PDF or a missing brief before the filing reaches the clerk’s desk, that is one fewer rejection to process, one fewer confused filer calling the clerk’s office, and one fewer delay in the case. Fewer bad filings in means less work on the back end.

And there is the access-to-justice angle. Fee waiver tracking, self-represented litigant completion rates, filing volumes by case type and county: this is the kind of data that grant applications and legislative testimony run on. A unified system produces it as a byproduct of doing its job.

Open source and what it means here

MUEFS is released under the AGPL-3.0 license. That means anyone can read the code, run it, change it, and contribute to it. If the state, a court, a legal aid group, or a law school wanted to pick it up and build on it, they could. No licensing fees. No vendor talks. No lock-in.

This is not an abstract idea. Government-built open-source software has a record. The most cited example is VistA, the electronic health records system the VA built and ran for decades. VistA achieved a 99.997% pharmacy prescription accuracy rate, was rated the top EHR in the country by physicians in independent surveys, and won Harvard’s Innovations in American Government Award.⁷ The VA released the code as public domain software. Jordan adopted it across its public hospital system. Several U.S. states used it in non-VA hospitals. One agency built it. The whole country benefited.

Closer to home, the federal judiciary is rebuilding its own filing infrastructure. The Administrative Office of the U.S. Courts is building a replacement for CM/ECF, the system that handles all federal court e-filing, with initial components in testing at six federal courts as of March 2026.⁸ If the federal courts can rebuild the system that runs their dockets, there is no technical reason Michigan cannot do the same for its state courts.

That said, open source is not free in the way people think. Someone still needs to host the system, maintain it, audit it for security, and patch it. Courts that adopted MUEFS or something like it would need IT support, in-house or on contract. The savings come from cutting per-transaction licensing fees and customizing without asking a vendor for a change order.

What comes next — and why this is realistic now

Michigan does not need another decade of incremental improvements to a fragmented system. The legislative foundation for a unified statewide platform was laid in 2015, when the Legislature passed six public acts that created a dedicated funding mechanism through modest increases in civil case initiation fees.⁹ The technical barriers to a single, modern, statewide e-filing platform have fallen. A unified front end with intelligent guidance, public search, high upload limits, and strong clerk tools can sit in front of existing case management systems through adapter layers.

Even in the courts that have gone live, meaningful improvements to the day-to-day filing experience have been slow to arrive.¹⁰ Core pain points that litigators encounter constantly — arbitrary upload limits that force splitting of large productions, the absence of intelligent prompts or validation for companion documents such as briefs and proposed orders, and the lack of built-in guidance on meet-and-confer certifications — have persisted for years. These are not minor inconveniences. They are friction that a modern e-filing system should eliminate by default.

What makes rapid progress newly realistic is the emergence of capable agentic coding assistants. These tools can understand large codebases, translate natural-language requirements into working features, generate tests, and iterate quickly under human direction. For a domain as rules-driven as Michigan Court Rules compliance, document validation, and structured workflows, agentic AI is exceptionally well suited. It can accelerate the encoding of clear procedural requirements, the building of smart prompts, and the maintenance of the system far more efficiently than traditional development alone.

This represents a more grounded and achievable contribution of artificial intelligence to access to justice than many of the more speculative applications currently discussed. Large language models that attempt to provide legal advice or draft case-specific arguments carry well-known risks of hallucination and incorrect statements of law. In contrast, using agentic tools to build and continuously improve court infrastructure — encoding the Michigan Court Rules into validation logic, guiding users through required steps, maintaining accurate public records, and surfacing the right documents at the right time — operates in a domain where outputs can be tested, reviewed by lawyers and clerks, and verified against objective rules. The result is better tools for everyone who uses the courts, not an attempt to automate legal judgment itself.

The combination of open-source foundations and modern development tools means that a single, high-quality e-filing platform for Michigan is no longer an inevitable multi-year vendor project. It is a realistic goal that the State Court Administrative Office, individual courts, legal aid organizations, or even groups of practitioners could pursue and improve over time.

A live, fully functional demonstration is available today at demo.tomcedoz.com — no download required. Anyone can walk through the actual filing flows, companion document prompts, clerk review queue, and public case search. The full source code is public at github.com/Tzodec1526/MUEFS under the AGPL-3.0 license. Anyone can clone it, inspect how it works, or build on it.

The worst outcome is that the current patchwork continues indefinitely. The best is that Michigan seizes the tools now available and delivers the unified, modern e-filing experience its litigators and citizens deserve.

Somewhere soon in Michigan, another lawyer will be offered the same two choices: drive there or mail it. There is now a third answer. Michigan just has to choose it.

---

Tom Cedoz is a partner at Husch Blackwell LLP in Michigan, where he focuses on commercial litigation and labor and employment matters, with substantial work at the intersection of artificial intelligence and the law. This article appeared in the Spring 2026 issue of the Michigan Litigation Journal, a publication of the State Bar of Michigan Litigation Section.

Endnotes

1. MiFILE is Michigan’s statewide e-filing portal, administered by the State Court Administrative Office (SCAO). As of June 2026, the official MiFILE portal states that it currently supports 127 trial courts. See mifile.courts.michigan.gov/availablecourts. ↩
2. The MUEFS source code is available at github.com/Tzodec1526/MUEFS under the GNU Affero General Public License v3.0 (AGPL-3.0). ↩
3. MCR 1.109 governs electronic filing and service in Michigan courts. It defines an electronic document largely by readability and permits several formats; the expectation that filings be in a text-searchable format comes primarily from SCAO’s MiFILE e-filing standards rather than from the rule text itself. ↩
4. MCR 2.119 requires that a motion presenting an issue of law include a brief citing supporting authority and provides for proposed orders. ↩
5. MCR 2.313 and related provisions require that a motion to compel discovery generally include a certification that the movant has in good faith conferred or attempted to confer with the opposing party. ↩
6. MCR 2.002 provides for waiver of fees and costs for indigent parties. ↩
7. The VA’s VistA system was released as public domain software under the Freedom of Information Act and received the Innovations in American Government Award from Harvard’s Ash Institute in 2006. See Press Release, U.S. Department of Veterans Affairs, VA Receives 2006 Innovations in Government Award (July 2006). ↩
8. Administrative Office of the U.S. Courts, Judges Outline Accelerated Modernization of Case Management System (March 10, 2026). ↩
9. See “Michigan’s E-filing Law: A Model Act,” Michigan Bar Journal (2016), discussing the six public acts effective January 1, 2016, that established the statewide e-filing framework and the Judicial Electronic Filing Fund financed by modest increases in civil case initiation fees. ↩
10. MiFILE integrates TrueFiling for the filer-facing portal with Hyland OnBase for document management and clerk workflow processing in the standard solution. See SCAO MiFILE resources and the MiFILE E-Filing Integration Guide. ↩