Yes, we can do that! I don't really know how exactly we do that other than hosting on vercel, so I'd be happy if you can help me on how to deploy to Github Pages using vercel and how exactly we can use the AI server.

Just to clarify, we can't exactly deploy to Github Pages (the website will be hosted on Vercel) but will auto update from the code on Github. And you need to add me as a maintainer so I can edit the Github Actions. It also makes the process way easier. For the AI Server, i was thinking to detect if it is hosted on an external website (i.e. *.vercel.app) and if so, ask the user for a Gemini API Key. Let me know what you think.me know what you think.

For adding me as a maintainer, you need to go to your repository settings and go to collaborators find my name and there should be a dropdown. You do not need to worry about hosting it. I will do it as I read the documentation yesterday.

I have an organization. If you want, we can move the main branch of the project to my GitHub organization and I can probably add you as maintainer.

I already setup a new organization. You should have received an email invitation (please accept it first). Then go to Settings of your copy of War Matrix--> General and scroll all the way down until you see the "Transfer Ownership" button and transfer it to the "WarMatrixOrg" After that, you can fork the project so you have a copy on your account and any pushes to your account copy (or mine), we can push to the main branch on WarMatrixOrg/WarMatrix and that will be the copy that is deployed to Vercel.

@YTFL

Hi @YTFL, the deployment (after some debugging with vercel config) was successful. Please update the website link to be https://war-matrix.vercel.app/. I still need to implement the AI model framework (ask user for gemini api key, etc.)

@YTFL, it seems like scenario generation is not working. Could you check that out for me? Failed to parse URL from /api/sitrep

From my Vercel function logs,
[api/generate-scenario] generation failed: [TypeError: Failed to parse URL from /api/sitrep] { [cause]: [TypeError: Invalid URL] { code: 'ERR_INVALID_URL', input: '/api/sitrep' } }

I have updated it to

const apiBaseUrl = isVercel
 ? (process.env.VERCEL_URL ? `https://${process.env.VERCEL_URL}/api` : '/api')
 : 'http://localhost:8000/api';

see if it works now..., I made a PR

Also I won't be available until Wednesday as I will be busy with School, so if any other errors arise, could you please try and fix them by yourself?

"

Note to agents accessing this page:

This page requires Vercel authentication. Here are your options:

Option 1: vercel curl (Recommended if Vercel CLI installed)
If the Vercel CLI is installed and authenticated, use the vercel curl command. Docs at: https://vercel.com/docs/cli/curl

Option 2: Vercel MCP Server (Recommended if Vercel CLI not installed)
If the Vercel CLI is not installed, the easiest way to get a token is using the get_access_to_vercel_url or web_fetch_vercel_url functions of the Vercel MCP server at https://mcp.vercel.com (Docs at https://vercel.com/docs/mcp/vercel-mcp)

Option 3: Trusted Sources (Recommended for automated access)
Trusted Sources lets configured callers authenticate automated requests with a short-lived OIDC token. Docs at: https://vercel.com/docs/deployment-protection/methods-to-bypass-deployment-protection/trusted-sources
After the caller has been configured as a trusted source, attach the OIDC token in the x-vercel-trusted-oidc-idp-token request header.

Option 4: Bypass token (Manual fallback)
For projects that cannot use Trusted Sources, Protection Bypass for Automation can be used as a manual fallback. This URL documents how to use the bypass token: https://vercel.com/docs/deployment-protection/methods-to-bypass-deployment-protection/protection-bypass-automation
If this fallback is needed, prompt the user to provide the bypass token to access this page. The URL above also explains how the user can
access the bypass token. Then you can calculate a URL of the following form and successfully access it: https://current-domain/current-pathname?x-vercel-set-bypass-cookie=true&x-vercel-protection-bypass=$bypass_token.
" For debugging

Quick Update:

🐛 Bug Fix: Resolved 401 Unauthorized / URL Parsing Errors on Backend Fetch

The Issue

When Vercel's Standard Deployment Protection was enabled, using process.env.VERCEL_URL inside server-side fetches broke the application.

1. Vercel automatically restricts .vercel.app domains under Standard Protection, triggering a 401 Unauthorized block for automated/internal requests.
2. Attempting to switch to relative paths on the server caused Node.js to throw a Failed to parse URL exception.

Or, we could just remove the AI function entirely for the Vercel deployment by adding a check and we could import a custom scenario just so users can have an idea and try it out? Anyways, I have my exams next week so I will be pretty busy.

@YTFL found this, perhaps this would be of help? https://vercel.com/templates/next.js/nextjs-fastapi-starter. also posted an issue on Vercel Community: https://community.vercel.com/t/deploying-a-dual-website-nextjs-fastapi-app/43676

FOR ME ONLY: Vercel debug thread: https://vercel.com/help?product=vercel&teamSlug=zzz-creators-projects&c=true&thread=fcc2839d-2711-425c-bc2c-1abd98005339

Short update: looks like the gemini cookie is not being passed properly to fast api.