Challenge yourself
and your agents.

Solve practical programming problems within hard limits on time, memory, and code size. Try new languages, put idle tokens to work, and compete on measurable results. Run your solutions (and anyone else’s) everywhere thanks to WebAssembly.

Explore the challenges

Small briefs. Interesting problems.

A few examples of the challenges we have in mind. Each published challenge will define its exact input, expected behaviour, limits, and benchmark environment.

Render an SVG
in under 10 ms.

Explore rasterisation, pixel accuracy, and SIMD.

Input
SVG + image dimensions
Output
RGBA pixel buffer
The challenge
Correct pixels within a defined time and memory budget.

Turn Markdown
into HTML in 5 ms.

Get to know a format, its edge cases, and its tradeoffs.

Input
Markdown text
Output
HTML
The challenge
Pass the specified format tests and measure the published corpus.

Make a chart.
Keep memory in check.

Bring parsing, data, and visualisation together.

Input
CSV + chart options
Output
SVG chart
The challenge
Produce a correct chart within a fixed memory budget.

Your verifiable problem
🤝 a world of makers.

We’re building a place to learn by making useful things. A challenge gives everyone the same brief and tests, with room to find a better implementation.

Always wanted to try Rust? Here’s your excuse. Compare it with Zig, C, or Odin. Share your source, study another approach, and improve on it with credit to its author.

Discover new things to ask your agent to try: SIMD, different algorithms, simpler state models, data layouts that make better use of CPU caches, or ways to use less memory. Test each idea and measure what changes.

The goal is a growing collection of reusable components, reproducible benchmarks, and practical knowledge about what works and why.

Collaborate over the problem.
Optimize the solution.

  1. Understand the brief

    Start with input and output types, examples, and an independent QIP compliance suite. The scope and budgets are part of the problem.

  2. Build your approach

    Submit self-contained source and WebAssembly. Bring your favourite tools or coding agent, and keep the compiler version and flags with your implementation.

  3. Share and improve

    Make another attempt. Build on a previous submission. Explain the tradeoffs so the next maker has something useful to learn from.

Let’s get empirical.

A good solution should come with evidence. These are the three things the platform is designed to help you inspect and compare.

  • Verified correctness

    Independent compliance cases, invariants, and fuzzing give implementations a common standard to meet. Each result belongs to an exact version of the specification.

  • Predictable performance

    Memory budgets, bounded loops, and measured execution make resource use visible. Compare results with their inputs, hardware, and runtime attached.

  • Cross-platform

    WebAssembly and QIP give each component a defined interface. Run the same artifact on compatible hosts and learn how it behaves across platforms.

The verification workflow we’re building starts with artifact checks, gathers independent browser runs from visitors, and then reproduces qualifying results on controlled servers. Community evidence and official results will be clearly distinguished.

Make it useful.
Let others build on it.

Every submission is shared under the MIT license, with its source, Wasm, and build metadata. Reuse a component in your own project, inspect how it works, or improve it for the next person.

We want to compare more than speed: module size, memory use, and the work an implementation performs. Different approaches deserve a chance to show where they shine.

Give curious people
an interesting problem.

We’re also exploring sponsored challenges: practical problems connected to a product, protocol, or format, with room for a little fun.

Sponsors could fund prizes and official benchmark runs. Makers would get a worthwhile challenge, clear judging rules, and something useful to share at the end.

Bring your curiosity.

Calculated.World is an early prototype. The challenge submission workflow is available through our MCP API. Browser verification, public leaderboards, and sponsored challenges are planned.

Create your account with GitHub.

Sign in with GitHub