Try it

npx github:VIABLE-Learning-Lab/viable

Or, in a clone, viable-engine try, with a package or an agent file if you have one. It opens a page on your own computer with six example courses (or what you named). They are the real learner pages, with one thing added beside them: a panel called What the AI did.

Ask the fractions tutor "just tell me the answer" and watch. You see what the teacher told the AI, the draft, the reviewer striking it with its reason, the rewrite, and what reached the learner. Change how far the AI may help, ask again, and the reply changes.

Demo mode, and your own model

With no key it runs in demo mode: the model's words are scripted, and everything around them is the real harness. No model is called, and what you type stays on your computer. With OPENAI_API_KEY or OPENAI_BASE_URL set, it runs on your model, and says so. --demo keeps it scripted anyway. The examples name VIABLE's models, so add --model <id> if your endpoint doesn't have them.

On your own model, a package with misconceptions to listen for costs one more call per message: it asks the model which of them the message shows. Learners' names and identifiers are never sent.

It answers on this computer only: it binds to 127.0.0.1, refuses any other name, and refuses a post from another website, so a page you are visiting can't change its settings, read what the AI did or spend your key. What you change in a try-out lasts until it stops. In demo mode a package's Fetch URL steps and http components stop too, so nothing leaves. Don't put a try-out behind a proxy: what the AI did is granted to whoever looks like this computer, and only a proxy that says it is one can be told apart.

What's on the page

• The panel. The teacher's settings, then each message as it happens. A small control changes how far the AI may help, and the goal, for this try only.
• The teacher's page (/teacher), with a made-up class from examples/class/. Accepting a suggestion changes the try-out's settings too. With your own package it shows what happened in the try-out.
• Resources (/resources): the catalog, the connected ones first, with what each does.
• Take it live: the commands from viable-engine deploy, and a prompt to give an AI assistant.

The density lab's frame loads from developers.viablelab.org, and its card says so. Nothing else on these pages loads from anywhere else.

A static copy for a website

viable-engine try --export site --base /try

This writes the same pages, with a small script answering their requests in the visitor's browser from the scripted replies, using the real engine. Put the folder on any static host. --base is the path it will be served under; leave it out for a site's root. Nothing a visitor types leaves their browser. The script is written only into a copy: no server route serves it.

A copy publishes what it runs on. The real engine answers in the visitor's browser, so all of it is in a file anyone can read: each agent's instructions, every question with its answer key, hints and explanations, and what the package holds for its knowledge bases. The bundled examples are public already. Anything else is refused unless you add --publish-answers, and even then a teacher's notes (watch_for) and the package's notes are left out. A course with a real quiz doesn't belong in one: use viable-engine deploy.

Give it its own path or subdomain, and not an origin with accounts. A static page can't have a fresh nonce per request, so its pages carry a fixed one and allow scripts from their own origin (script-src 'self'). On an origin where people sign in, those pages are a weaker neighbor than the site deserves. The tour there clears only the copy's own storage (viable-chat: and viable-try:).

The panel on your own server

It is off by default. It carries instructions, drafts and reviewers' reasons, so a learner must never get it. The route decides on each request, never the page.

viable-engine serve ... --panel grants it to whoever holds the token. With no --token-env, the server makes a token at start and prints it once, so the grant is never decided by a header. Sign in at /teacher; its links open a page with ?panel=1, which is what asks for the panel. A page without it stays a learner's page, even in the browser you signed in on.

That browser is the thing to watch on a shared computer. The session ends when the browser closes, or after 8 hours, and the panel and the teacher's page each have a Sign out: use it before a learner sits down. Until then, anyone at that browser who types ?panel=1 sees what you see. In code, learnerRoutes({ traces, panel }); see src/learner/watch.js, where watchSlot is what a page gets and watchHooks is what a run tells it.

Test results as a page

viable-engine test agent.json cases.json --judge-model <id> --html

This serves the results on this computer until you press Ctrl-C. Add --out report.html for one file instead.

This page comes from the open repository: engine/docs/try.md. To change it, change it there.