Open-source platform for 3D AI agents. Turn text or a photo into a rigged, animated GLB avatar, give it an LLM brain, memory and a wallet, and embed it anywhere with one web component. Remote MCP server, x402 payments, Three.js, browser-native.
JavaScript 3d-generationai-agentsanimationavatardigital-humangenerative-aiglbgltfimage-to-3dllmmcpmcp-servermodel-context-protocolsolanatext-to-3dthreejsweb-componentswebglwebxrx402
/**
* Tutorial library invariants.
*
* A tutorial is only real when three things agree: the markdown at
* docs/tutorials/<slug>.md (the content), an entry in
* public/tutorials-manifest.js (how anyone FINDS it), and a route that
* renders it. The manifest is loaded as a classic script by both /tutorials
* and /tutorials/<slug>, so a file with no entry is written, shipped, and
* unreachable from navigation: it exists only for whoever already knows the
* URL. That drift is invisible in review and has happened repeatedly, which is
* why it is pinned here rather than left to a periodic sweep.
*
* The manifest is a browser script (an IIFE assigning window.TUTORIALS), not a
* module, so it is evaluated in a tiny window shim instead of imported.
*/
import { readFileSync, readdirSync, existsSync } from 'node:fs';
import { resolve } from 'node:path';
import { describe, it, expect } from 'vitest';
const root = resolve(__dirname, '..');
/** Evaluate the manifest IIFE the way a browser does and return window.TUTORIALS. */
function loadManifest() {
const source = readFileSync(resolve(root, 'public/tutorials-manifest.js'), 'utf8');
const win = {};
new Function('window', source)(win);
return win.TUTORIALS;
}
const TUTORIALS = loadManifest();
const TIERS = ['easy', 'middle', 'advanced'];
const markdownSlugs = readdirSync(resolve(root, 'docs/tutorials'))
.filter((f) => f.endsWith('.md'))
.map((f) => f.replace(/\.md$/, ''));
describe('tutorials manifest', () => {
it('parses into a non-trivial list', () => {
expect(Array.isArray(TUTORIALS)).toBe(true);
expect(TUTORIALS.length).toBeGreaterThanOrEqual(50);
});
it('every entry carries the fields both templates render', () => {
for (const t of TUTORIALS) {
expect(typeof t.slug, JSON.stringify(t)).toBe('string');
expect(t.slug).toMatch(/^[a-z0-9-]+$/);
expect(TIERS, `${t.slug} tier`).toContain(t.tier);
for (const field of ['title', 'blurb', 'builds', 'time']) {
expect(typeof t[field], `${t.slug}.${field}`).toBe('string');
expect(t[field].length, `${t.slug}.${field}`).toBeGreaterThan(0);
}
expect(typeof t.ctaPrimary?.label, `${t.slug}.ctaPrimary.label`).toBe('string');
expect(typeof t.ctaPrimary?.href, `${t.slug}.ctaPrimary.href`).toBe('string');
}
});
it('has no duplicate slugs', () => {
const seen = new Set();
const dupes = [];
for (const t of TUTORIALS) {
if (seen.has(t.slug)) dupes.push(t.slug);
seen.add(t.slug);
}
// A duplicate renders the same card twice and makes the prev/next pager
// in /tutorials/<slug> jump backwards.
expect(dupes).toEqual([]);
});
});
describe('manifest and markdown agree', () => {
it('every manifest entry has the markdown its viewer fetches', () => {
const missing = TUTORIALS.filter(
(t) => !existsSync(resolve(root, 'docs/tutorials', `${t.slug}.md`)),
).map((t) => t.slug);
expect(missing, 'manifest entries with no docs/tutorials/<slug>.md').toEqual([]);
});
it('every tutorial markdown file is listed in the manifest', () => {
const listed = new Set(TUTORIALS.map((t) => t.slug));
const orphans = markdownSlugs.filter((slug) => !listed.has(slug));
// An orphan is written, shipped, and unreachable: /tutorials never links
// it and the pager skips it. Add a manifest entry when you add the file.
expect(orphans, 'tutorials with no manifest entry (unreachable from /tutorials)').toEqual([]);
});
});
describe('preview models and CTAs resolve', () => {
it('every previewModel points at a GLB that ships', () => {
const broken = TUTORIALS.filter((t) => t.previewModel).filter(
(t) => !existsSync(resolve(root, 'public', t.previewModel.replace(/^\//, ''))),
).map((t) => `${t.slug} -> ${t.previewModel}`);
expect(broken, 'previewModel files missing from public/').toEqual([]);
});
it('every internal CTA href is a routable path, not a bare guess', () => {
const routes = JSON.parse(readFileSync(resolve(root, 'vercel.json'), 'utf8')).routes;
const pages = readFileSync(resolve(root, 'data/pages.json'), 'utf8');
const routable = (href) => {
const path = href.split(/[?#]/)[0].replace(/\/$/, '') || '/';
if (pages.includes(`"${path}"`)) return true;
// Docs pages are served by the docs SPA, not listed one by one.
if (path.startsWith('/docs/')) {
return existsSync(resolve(root, 'docs', `${path.slice('/docs/'.length)}.md`));
}
return routes.some((r) => {
if (!r.src) return false;
try {
return new RegExp(`^${r.src}$`).test(path);
} catch {
return false;
}
});
};
const broken = TUTORIALS.map((t) => t.ctaPrimary)
.filter((cta) => cta.href.startsWith('/'))
.filter((cta) => !routable(cta.href))
.map((cta) => cta.href);
expect([...new Set(broken)], 'CTA hrefs that match no page or route').toEqual([]);
});
});