Skip to content
ImageLurk Lab

A verifiable foundation for AI production.

Lurk Lab is building a stack for AI agents to compute privately, act under explicit authority, and produce work others can independently verify.

Our work begins with Lurk, a general-purpose language for verifiable computation, and extends through the systems that let independent agents build on one another’s work.

Lurk

Lurk is a Turing-complete, statically scoped dialect of Lisp, influenced by Scheme and Common Lisp. Lurk program execution can be proved in zero-knowledge.

Programs and data are content-addressed. Functions can receive and return other functions.

Lurk expressionEditable
;; (FIB TARGET) computes the element of the Fibonacci sequence at TARGET (zero-indexed).
(letrec ((next (lambda (a b n target)
                 (if (eq n target)
                     a
                     (next b
                           (+ a b)
                           (+ 1 n)
                           target))))
         (fib (next 0 1 0)))
        (fib 1))
Evaluation result

Verifiable computation

Lurk expressions can be proved, allowing for verifiable outsourcing of computation. This demonstration walks through evaluation, proving, and server verification.

Proof expressionEditable
((lambda (a b)
    (+ (* a 3) b))
  9 7)
Proof evaluation result

Latest posts

  1. LurkScript, Concurrent Lurk, and LurkOS: Unlocking the Next Era of ZK‑Powered Applications

  2. Lookups in Lurk: Part 2

  3. Lurk 0.5 Benchmarks

  4. Lookups in Lurk: Part 1

View all posts →