Hosting the website

The CopilotForge website is built from this repository, so the site, the catalogue and the plugins cannot drift apart (decision D14). It is designed for professional web hosting at https://copilotforge.ai/.

Part Built by Where

Core pages (home, catalogue, programmes, templates, programme builder, author, contributors, privacy) scripts/build-site.py site/ (committed and checked)

The full site: every guide, department, skill, template and programme page, downloads, licence texts, 404, .htaccess, robots.txt, sitemap.xml scripts/build-hosting.py dist/website/ and dist/copilotforge-website.zip (git-ignored)

Part	Built by	Where
Core pages (home, catalogue, programmes, templates, programme builder, author, contributors, privacy)	`scripts/build-site.py`	`site/` (committed and checked)
The full site: every guide, department, skill, template and programme page, downloads, licence texts, 404, `.htaccess`, `robots.txt`, `sitemap.xml`	`scripts/build-hosting.py`	`dist/website/` and `dist/copilotforge-website.zip` (git-ignored)

Build the hosted site

bash scripts/build-all.sh
python3 scripts/build-hosting.py

The build checks that every link and image on every page resolves and fails if one does not. Pass --site-url to build for another address.

Like the D365Forge website, every reference to the GitHub repository is a link: the GitHub item in the navigation, the "View on GitHub" button, links to files that are only in the repository, and every repository address written out in the text (decision D24). While the repository is private (open decision O5), only people who have been given access can open those links; everyone else sees GitHub's "page not found". To show repository-only references as plain text instead, set GITHUB_LINKS = False in scripts/_site.py, or pass --no-github-links, and rebuild. The build fails if a repository address is shown on any page without being a link.

Upload to your web host

In your hosting control panel, open File Manager for the copilotforge.ai website.
Open public_html. If it holds a placeholder page, such as default.php, delete it.
Upload copilotforge-website.zip into public_html.
Extract it into public_html itself, not into a new folder. The archive holds the site files at its top level, so index.html ends up directly in public_html.
Delete the archive once it has been extracted.
Turn on SSL for the domain and force HTTPS if it is not already on. The .htaccess file also redirects HTTP to HTTPS.
Open the site and check the home page, a skill page, the programme builder and a download.

The site is plain HTML and CSS with two small scripts (the catalogue search and the programme builder). It needs no PHP, no database and no build step on the server.

Update the hosted site

Rebuild, then upload and extract the new archive over the old files. If pages have been removed or renamed, empty public_html first so that stale pages do not linger.

Privacy

The privacy page says the site sets no cookies and has no analytics, and that the web host may log requests. If you add analytics, a cookie banner or a contact form, update the privacy page in scripts/build-site.py first.