UDIF: The Universal Data Interchange Format
Invented by Angela Benton
What It Is
UDIF, the Universal Data Interchange Format, is an open data standard designed to make your data truly portable. Not portable in the way platforms use that word, meaning transferable between their systems on their terms. Portable in the literal sense: a file format you own, that lives where you choose, that any system can read without anyone's permission.
I invented UDIF because the data economy had a structural problem that no platform was incentivized to solve. The people generating the most valuable data — users, creators, consumers — had no format that belonged to them. Every platform held your data in its own proprietary structure, on its own servers, under its own terms. You couldn't leave without losing everything you'd built. That's not an accident, it's the architecture.
UDIF was designed to break that architecture.
The Origin Story
I built UDIF at Streamlytics, a consumer data company I founded in 2018. At Streamlytics we processed over ten billion data points from across a hundred and twenty-five countries; streaming behavior, social data, and consumption patterns. We were giving people access to their own data at a scale that hadn't been done before.
And the more we built, the more clearly I saw the problem. Even when users could access their data, they couldn't do anything meaningful with it. There was no standard format. No way to carry it from one system to another. No way to prove it was theirs. The data existed, but the infrastructure to make it sovereign didn't.
That's when I started designing UDIF; a file format that could hold user data in a way that was platform-agnostic, human-readable, and verifiable. Not locked inside any company's database. Not dependent on any company's cooperation to access. A file that lived with the person who generated it.
I filed a patent on the underlying system in 2021. The patent application (US20220300636A1) explicitly describes UDIF and its architecture for standardizing data across platforms. The work was covered by Inc. Magazine, MarTech Series, GlobeNewswire, and others as part of Streamlytics' broader mission to give people control over their own data.
What UDIF Does
At its core, UDIF solves a format problem.
When your data lives inside a platform's infrastructure, it exists in that platform's proprietary format. Moving it requires that platform's cooperation, their export tools, their timeline, their terms. When the company changes its policy, gets acquired, or shuts down, your data's accessibility changes with it. You never actually had a file. You had access to a file that someone else owned.
UDIF gives you the file.
The standard defines how data should be structured so that it's readable by any system, storable anywhere, and verifiable as authentic. It's the data equivalent of a PDF — a format that works regardless of what application you're using to open it, that you can save to your own device, that you can share without losing control of the original.
The difference sounds simple. The implications are significant. A file you own can't be held hostage. It can't be deleted by a platform policy change. It can't be used to train a model without your consent. It can't be used as a switching cost to keep you locked into a system you'd otherwise leave.
UDIF 2.0
The original UDIF standard was built for the data economy that existed at Streamlytics — streaming behavior, social data, consumption patterns. The problem it solved was real and the architecture was sound, but the data landscape has shifted fundamentally.
The most valuable data people generate now isn't their streaming history. It's their cognition. Their AI interactions. The accumulated record of how they think, what they know, what they care about, how they work. Every conversation with an AI system is an act of intellectual output that has value — and right now, all of that value lives on platform servers, in platform formats, under platform terms.
UDIF 2.0 extends the original standard into the AI interaction layer. It defines how your AI context — your conversations, your preferences, your intellectual history, your values, your ways of thinking — can be structured in a format you own and control. A file that lives on your device or your own personal cloud. That any AI system can read. That moves with you rather than staying behind when you switch platforms.
The Heirloom protocol builds on top of UDIF 2.0 by adding cryptographic provenance — a verifiable record that your data originated from you, is authentic, and hasn't been altered. So it's not just portable. It's provably yours.
Why This Matters Now
In March 2026, Anthropic launched a memory import feature allowing users to move their data from ChatGPT into Claude. Google launched a similar feature for Gemini. Both were called portability.
Neither of them is.
What they built are migration tools — ways to move your data from one company's servers to another company's servers. When the transfer completes, your memories live in the new platform's infrastructure, encrypted by them, governed by their terms, subject to whatever policy decisions they make next year. You changed landlords. You didn't buy the building.
Real portability requires a standard like UDIF — a format that isn't owned by any platform, that any system can read, that allows your data to live in infrastructure you control. A personal website. Your own cloud environment. A sovereign data layer that you connect to AI tools as needed, on your terms, rather than one that lives inside any single company's ecosystem.
This is the architecture the data economy has needed since the beginning. The AI layer makes it urgent.
The Bigger Thesis
UDIF is one piece of a larger argument I've been making for nearly a decade: that the infrastructure question is the sovereignty question. You cannot own your data if you don't own the format it lives in. You cannot own the format if you don't own the infrastructure it runs on. And you cannot own the infrastructure if you're entirely dependent on platforms for access to your own digital life.
The New Internet — the one worth building — is one where individuals have genuine sovereignty over their data, their identity, their creative output, and their AI context. Where the context layer belongs to the person who generated it, not to the platform that processed it. Where your intellectual history is an asset you own, not a switching cost someone else holds.
UDIF is the standard that makes that possible. Heirloom is the protocol being built on top of it.
Documentation & Sources
US Patent Application US20220300636A1 — System and Method for Standardizing Data — View on Google Patents
PCT Application WO2022198036A1 — View on Google Patents
MarTech Series / GlobeNewswire (March 2022) — Streamlytics Enables Public to Attach Real-World Data to NFTs
Inc. Magazine — Angela Benton: Streamlytics Developing a Universal Standard to Help You Manage Your Data
Continue Reading
→ Anthropic and Google Just Proved My Thesis (essay) → What Heirloom Is Building → Watch the Working Session Replay
Angela Benton is a three-time founder and the inventor of UDIF. She previously founded NewME, the first accelerator for underrepresented founders in Silicon Valley, and Streamlytics, where she built one of the first consumer data portability platforms at scale. She is the founder of Heirloom and an advocate for digital sovereignty and human-centered AI.