MCFN-OS gets a real build start
Auddia says it has crossed from talk to construction on the McCarthy Finney Operating System, or MCFN-OS. In a June 4, 2026 Globe Newswire release carried by Benzinga, the company says it engaged Chase J. Thompson, founder of AI integration firm Arqa, as Platform Architect.
The stated milestone is simple but meaningful. Thompson has begun work on the platform, marking MCFN-OS’s transition from conceptual architecture to an active, in-progress build, according to the Benzinga-linked release.
The platform pitch: “agents do the work, people do the approvals”
Thompson’s explanation targets a specific failure mode in many business AI tools: they act like systems of record rather than systems where work happens. In the release, he contrasts traditional tools that capture output after the fact with MCFN-OS as “the layer the work actually happens in.”
He describes a chat-first operating model where users direct work in plain language and agents carry it out, from drafting a press release to coordinating workflows across subsidiaries. The company is also explicit about governance. Thompson says the design keeps agents from acting without human oversight. He claims the system includes a unified substrate and a real audit trail. Everything produced traces back to a source, and approvals and edits get recorded.
He also frames the security posture as operational discipline for a public company. In his words, the system is “deliberately not autonomous,” and “nothing leaves the building without a person approving it,” per the Benzinga-linked release.
That matters because agentic systems often fail in the boring places. Approval loops, logging, and provenance are where teams either win or get dragged into compliance chaos. MCFN-OS’s pitch is basically that those controls live inside the platform layer, not on the desks of the people running the business.
What ships first and when
Auddia’s release says the first production module is targeted for delivery in Q3 2026. It repeats that each module should build on the same foundation: a shared institutional memory and a common agent runtime.
It also names components that aim to support reuse across functions and business units. The release mentions a knowledge graph plus a Web3 trust layer planned for MCFN-OS, along with capabilities like agentic AI execution, cross-vertical intelligence, workflow automation, and chat-first interfaces.
How the web3 layer is supposed to fit
The release uses “Web3-enabled” language in multiple places, but it keeps the specifics high level.
It says MCFN-OS will include “blockchain-anchored trust infrastructure,” plus “Web3-verified identity and provenance” and a planned Web3 trust layer. It also says each module will rely on a shared institutional memory and a common agent runtime that includes a knowledge graph.
For readers who care about what gets built versus what gets promised, the most concrete line is the governance detail Thompson gave. The less concrete part is what “blockchain-anchored trust” will actually cover in the first module.
Here’s what the release claims on the module timeline and platform components.
| Item | Claim from Benzinga-linked release |
|---|---|
| Platform builder step | Thompson engaged as Platform Architect, starting MCFN-OS build |
| First module | “working agentic system” with unified substrate and audit trail |
| Target delivery | Q3 2026 |
| Shared foundation | Shared institutional memory and common agent runtime |
| Planned data and trust pieces | Knowledge graph and Web3 trust layer |
| Interface model | Chat-first, no dashboards or software learning curve |
Merger context: why Auddia is building now
The build is tied to a proposed merger. Auddia says it entered a definitive merger agreement on February 17, 2026, combining Auddia Inc. with Thramann Holdings, LLC.
The release says the end result would rename Auddia to McCarthy Finney and move to trading under ticker MCFN. It also says McCarthy Finney will operate as an AI holding company for four portfolio companies: LT350, Influence Healthcare, Voyex, and Auddia.
Thramann Holdings is described as fully owning LT350, Influence Healthcare, and Voyex. The release describes LT350 as a distributed AI data center company. Influence Healthcare is described as a health-tech company using AI and blockchain. Voyex is listed, but without detail in the excerpt.
Auddia’s CEO, Jeff Thramann, says Chase’s engagement is the first tangible milestone toward the operating system that would support McCarthy Finney’s subsidiaries after closing.
The team plan: platform work, not app sprawl
Thompson is set to lead a lean Platform Group responsible for AI agent orchestration, knowledge graph engineering, Web3 identity and provenance, workflow automation, and application-layer interfaces, according to the release.
That structure lines up with the core promise in the platform pitch. The company is trying to avoid a future where each subsidiary gets its own agent stack, its own logging scheme, and its own provenance story. Thompson says the same system should extend across subsidiaries because it is built as a general operating layer, not a one-off.
For now, this is a milestone announcement. It does not provide code, third-party verification, or detailed threat models. Still, it does put a date on the first module and ties the agent governance story to a platform-wide audit trail, human approvals, and traceability claims made in the Benzinga-linked release.