Example apps

Sixteen working apps ship with Wardian. Each one teaches one thing well, and each one is a package you can open, read and copy. On the website, each one's page runs it in your browser when it needs no server, and every one can be downloaded and imported. They are built into Wardian and in the app list from the first start, however you installed it, and in apps/ in the repository.

Every example passes wardian check and a browser test (tests/run-examples-e2e.sh). Read an example's README.md first: it explains each file and how the data moves.

Which example shows what#

Modules#

adder#

The smallest app: a WebAssembly file with an add function, and nothing else. Wardian draws a card with two inputs and a Run button. Start here to see what Wardian does with no help at all. How it works · Source

unit-converter#

Thirty-two everyday conversions (temperature, length, mass, volume, speed, fuel use, pressure, power, food energy and body mass index) as plain numeric functions. There is no HTML and no JavaScript: Wardian draws one card per function, with an input per parameter, and the viewer can arrange the cards. The function names carry the units, such as km_to_miles and bmi_kg_m, because the host labels the inputs only by position. How it works · Source

number-lab#

Twenty functions on 64-bit whole numbers: primes (exact for every 64-bit number), factors, greatest common divisors, modular powers, Collatz steps, Fibonacci, factorials and binomials. Wardian passes i64 as JavaScript BigInts, so a 19-digit answer stays exact. A result of -1 means the input is out of range or the answer is too big. How it works · Source

Pages#

text-tools#

Counts words, sentences and reading time, lists the most common words, hashes the text with SHA-256 and changes its case, live as you type, paste or open a file. Rust does the work, with no crates. The page copies the text into the module's memory once per change, and gets structured results back as JSON text. A 2 MB paste updates in about a tenth of a second. How it works · Source

image-lab#

Open a photo and apply filters: blur, sharpen, edge detection, tone and posterize, with a live preview, a before-and-after slider, undo and an RGB histogram. The filters run in WebAssembly on one reused pixel buffer, and the histogram is counted there too. The page sends the photo nowhere; the only way out is the PNG you save. How it works · Source

life#

Conway's Game of Life on a grid whose edges wrap around, with the whole simulation in WebAssembly: two grids swapped in turn, and step(n) runs many generations in one call. Draw cells, load classic patterns such as the Gosper glider gun and the pulsar, and switch to the HighLife or Seeds rule. When you ask your system for reduced motion, it starts paused. How it works · Source

mandelbrot#

The Mandelbrot set, rendered by WebAssembly into one pixel buffer per frame. Click to zoom; save the view as a PNG. A 1600 × 1066 view at 400 iterations is hundreds of millions of steps, done in one call. How it works · Source

focus-timer#

A Pomodoro timer with focus, short and long breaks, a dial and a soft chime. When a focus session ends, it sends the session on the channel focus.session, if you allow it, and says why when it cannot. It is half of a pair with focus-log, and shows how a page uses /sdk/wardian.js. How it works · Source

Suites#

loan-planner#

Six sealed parts plan a loan: inputs, an engine in WebAssembly, a planner, a summary, a chart and a CSV export. The planner calls the engine once per change and sends one plan to every viewer. This is the reference suite. How it works · Source

habit-tracker#

Tick off your daily habits, see each one fill a year-long heat map, and keep your streaks going. One part with no panel owns the check-ins and the clock, another owns the list, and the rest ask them through methods and receive copies on retained topics. A JSON backup moves everything anywhere; a restore checks the whole file and asks first. The views move to the next day at midnight while the app is open. How it works · Source

csv-explorer#

Open a CSV file, or the sample, and explore it: sort, filter, page, see each column's statistics and a chart, and download the rows you kept. Five small parts share the work. The rows live in the app's own SQLite database, so a 100,000-row file loads in about a second, pages, sorts and filters in a fifth of a second, and is still there after a reload. Every value typed in the filter is a bound parameter, never SQL text. How it works · Source

meeting-notes#

Paste meeting notes or a transcript and get a summary, the decisions, an editable checklist of who does what, and the open questions, ready to export as Markdown. With a Claude provider set up, Claude reads the notes, with a Stop button and a clear message for each error code. Without one, a local reader does the job, and the app says how to turn AI on. How it works · Source

monte-carlo#

A schedule risk simulator. List a project's tasks with best, likely and worst estimates and what each one waits for, then run up to a million trials in a Web Worker: about 2.3 seconds, while the page still answers at once. See the finish at 50, 80 and 95 percent confidence, a histogram with an S-curve, and how often each task is on the critical path. Stop keeps the trials already done, and a visible seed makes every result repeatable. How it works · Source

focus-log#

Receives the sessions focus-timer sends, keeps them in its own database, and shows today against a goal, the last seven days, time per label and the full history. Wardian replays the latest message to a log that opens late; each session has its own key, so a replay never counts twice. Together the two show a channel from start to end, with your permission at each step. How it works · Source

usl-lab#

Fits the Universal Scalability Law to load-test results and shows where a system stops scaling. It runs its fitting engine in a worker, from code it reads with source, and receives tables from the Splunk table app. How it works · Source

splunk-table#

Runs a Splunk search on the server, loads up to a million rows into its database, and shows them a page at a time. It sends a reference to the table on a channel, so the USL lab can read the rows it needs. It needs a Splunk account. How it works · Source

Make your own#

Start from a template, or from an example that is close to what you want:

wardian new suite apps/my-tool
cp -R apps/csv-explorer apps/my-explorer

Then follow Building Wardian apps, and Designing a suite for a suite.

This page is docs/site/examples.md in the repository. Something wrong or missing? Change that file.