(Re) Introducing Stoffel: Announcing Stoffel 0.1.0

One morning, you wake up, doom scroll until you see yet another breach of a massively used online service. You think to yourself, "Shouldn't they have known better?" And to anyone who has built or designed consumer-facing products, you'd be right.

However, teams are perpetually under pressure to ship more, ship faster, ship bigger. New features take priority over security and privacy. When security and privacy do need to enter the picture, the minimum imposed by some bureaucratic regulator is done. This often means integrating data mapping software, often branded as "data privacy software", updating your privacy policy, and applying "anonymization" techniques to your data. Clearly, based on the widespread increase in breaches in 2026, these measures have done nothing to even minimize the loss in data and user trust.

The advent of AI and LLMs doesn't help either. Recent research has shown that even with precaution, LLMs are able to match pseudonyms to known identities of known users. So, even with anonymized data, when combined together, you can still get a profile of a person without needing to know what they'd consider to be highly personal infomation (e.g. name).

So where does that lead us as builders?

Stoffel

Stoffel equips builders with the mental model to design and engineer more privacy-first architectures for their apps without needing to rewrite the entire app like a jail cell.

What we are releasing

Stoffel 0.1.0 contains a full runtime for multiparty computation, a subfield of cryptography that concerns itself with how to transform single party programs into multiparty programs without leaking the inputs into those programs to all of the other parties. We've written about the basics of MPC computation in our introduction to secret sharing and beaver triples articles.

Under the hood, Stoffel 0.1.0 contains a full MPC runtime:

• Stoffel Lang for writing computations.

• Compiler and .stflb bytecode as a portable program boundary.

• Stoffel VM, a register-based runtime with MPC-oriented execution surfaces.

• Rust SDK for compile, load, and run workflows. This is what a developer actually integrates into their app

With Stoffel, you don't need to understand the nitty gritty of how MPC works. The Stoffel runtime handles those details for you.

The goal is not to add another policy layer on top of the same architecture. The goal is to make the private part of the computation something a developer can write, run locally to inspect it quickly and then deploy. Once deployed, using Stoffel via the Rust SDK is no different than using any other SDK.

Try it Now

Unlike previous privacy tools, the installation flow for Stoffel is familiar

$ curl -fsSL https://get.stoffelmpc.com | sh

$ curl -fsSL https://get.stoffelmpc.com | sh

$ curl -fsSL https://get.stoffelmpc.com | sh

Then all you have to do is

You can get the details in our Getting Started docs

What can you build with it today?

The Stoffel stack was built to ensure that you, while you do need to adjust your mental model a little, don't need to relearn how to program completely.

With Stoffel, you can build

• Threshold signatures

• Statistical algorithms

• Aggregate statistics

And if you are feeling a little lucky, you can even rebuild common operators from scratch but privacy first such as:

• Comparison algorithms

• Sorting algorithms

• Matrix and vector operations

You can check out our demo programs in our examples folder.

Who should try it

We want to hear from anybody building products with a system component that would benefit from increased privacy. Our wishlist of product categories:

• Key management services

• Product analytics

• Web analytics

• Graph analytics

• Certificate management

Even if you don't work on a product on our wishlist, we'd love to hear from you.

What's coming next?

As part of our mission to make private compute accessible to empower developers to write privacy-first apps, our next set of features are focused on making it easier for you to integrate into your apps. Here's a quick preview:

• Typescript, Javascript, Python and Go SDKs

• An expressive standard library for common functionality

• Faster MPC protocols

• More tutorials

Start here

The point is not that Stoffel makes privacy easy (He does make it easy though). The point is that developers need better privacy developer tools that already work how they do without burdening them.

Stoffel 0.1.0 is a step in that direction: a language, bytecode format, VM, CLI, and SDK for making private computation workflows concrete enough to run, inspect, and run.

If you try it, we want the rough edges. If the programming model is confusing, tell us. If the Rust SDK is missing an example you need, tell us. If your use case does not fit, tell us why.

• Build your first Stoffel app

• Read the release/repo

• Read the docs

• Bring us your app, use case, or rough edge

Acknowledgements

Thank you to the entire Stoffel Labs and HashCloak team (Gabe, Garry, Cordian and Hernan) for shipping this massive release. Thank you to our design partners for taking a risk on us despite our repeated delays. Thank you to our infrastructure partners for working with us. Thank you to Zellic and Veria Labs for being a part of our security posture.