Introducing Quint: A New Company for a New Era of Software Quint - Executable specs for reliable systems

Thank you to the community that has been with us from the start.

We're building something we believe in, and we're building it for you.

Read the full story: bit.ly/4tyzHTn

And something new is coming. Quint Studio.

Capture your intent as executable, high-level system descriptions with verifiable properties. Validate behavior during design, carry it into testing, and surface mismatches in production.

Sign up for early access: bit.ly/4c7z9gL

Quint, the language, stays open source, for everyone. Always.

Quint Executable specs for reliable systems

We've launched a brand new website and home for Quint.

quint.sh/

What if your spec wasn't just documentation? With Quint, specs come alive.

Simulate systems, verify correctness, and connect your intent directly to tests and code.
Continuous confidence, from design to deployment.

AI is generating more code than ever, and formal methods are finally becoming accessible. Quint sits exactly at that intersection. The timing has never been better.

Quint started as a question: how do we actually trust the systems we're building?

Today it becomes its own company, with @bugarela.com as CEO.

Two Bugs in the SAFE Predicate: Finding and Formally Verifying Liveness Failures in Byzantine Lattice Agreement Quint - Executable specs for reliable systems

Read the full walkthrough: bit.ly/4ejdwvi

We get a 27-step lasso counterexample in the original algorithm, and no violations once we include our proposed solution.

We found two liveness bugs in published GWTS pseudocode via simulation and reasoning. Then we used Quint’s new TLC integration to formally verify liveness as a temporal property across the full reachable state space.

Thanks for building and sharing. Tag us if you want to be in the next one.

Renan Liberato Sant'Ana modeled a concurrent fix in Quint before writing any code. Found two bugs in the LLM-suggested solution. One silently killed a feature for the entire user session. Ordering issues you'd never catch in code review.

bit.ly/4vmKKAa

@curlykoder is porting Quint-Connect to Zig. MBT as a native build.zig step, with realtime invariant violation detection and visualizations. Still early, and already finding bugs.
bit.ly/4vtEeYz

@twistin456 is building a Quint-to-Go MBT pipeline: write a Quint spec, generate ITF traces, parse and assert in Go. Once you have correct traces, implementation consistency across languages takes care of itself.

bit.ly/4c2HORB

@ohdearlordylord ported D&D 5th edition combat rules to Quint + XState. TTRPG rulebooks are state machines encoded in natural language and community Q&A is basically QA. LLMs translated the text, Quint tested the application against formal rules.

bit.ly/4mrHlfw

Four things the Quint community built and shared recently.

D&D combat rules, multi-level Go MBT pipelines, Quint Connect for Zig, and a race condition caught before a single line of code was written.

How to Write Inductive Invariants Quint - Executable specs for reliable systems

Check out our walkthrough: quint-lang.org/posts/induc...

"Inductive invariant" sounds intimidating. They are a powerful way of proving the absence of design bugs in distributed systems. Quint has tooling for them.

(happy april fools 🥸 actual violations still reported as violations)

Introducing Quint Vibes: model checking felt too harsh. Starting today, instead of invariant violations, Quint returns a vibe assessment.

🌟 immaculate vibes -- your spec is fine, probably
⚠️ off vibes -- something feels wrong but we can’t say what
💀 no vibes -- you have bigger problems

choreo/examples/atm_distributed_lock at main · informalsystems/choreo Choreograph distributed protocols in Quint. Contribute to informalsystems/choreo development by creating an account on GitHub.

Check out the example at: bit.ly/3NVkYC4

We have a new Quint example using Choreo: multiple ATM machines keep distributed locks to prevent overdraft. Includes model, properties, example runs for happy and unhappy paths, and several witnesses ✅

Ok, nice. Quint caught a design bug. Not a huge one, but at least the agent was able to use it to change its design.

The general thesis, I think is right. You can use formal methods as guardrails for agents to steer a design, in part, by themselves.

Modeling the Faerie-Gold Vulnerability Quint - Executable specs for reliable systems

In the breakdown: symbolic crypto modeling, simulating zk-SNARKs proofs as direct witness checks, and why the fix works.
bit.ly/4bGUFYh

St. Patrick's Day may be over but we're still thinking about fake gold 🍀

New post: we spec'd the Zerocash faerie-gold vulnerability in Quint and could clearly see the scenario where it gets exploited. Then we saw how a Zcash-like fix solves it.

Meet the software engineers who aren’t letting AI push them out As layoffs rise and AI reshapes the profession, some developers are doubling down on the craft and learning how to be indispensable.

Full Article: bit.ly/4sp2x81

@leaddev.com just published a piece featuring Quint's @bugarela.bsky.social on engineers who aren't letting AI push them out.

On why AI makes her work more relevant, not less:

"It actually feels like the industry is now catching up to the problem I've spent the past several years working on."

Checking Properties Quint - Executable specs for reliable systems

Read more about witnesses in the docs:
quint-lang.org/docs/checki...

Here's @bugarela walking through a realistic example of witnesses in practice, back when we were still writing Quint specs by hand.

x.com/bugarela/st...

Ever find yourself asking "how do I even get to the state I want to test?" Witnesses in Quint answer that question. Define the state you're looking for and Quint finds an execution path that gets you there. No manual trace construction required.