Skip to content

USER_INSTRUCTIONS.md

Latest commit

 

History

History
100 lines (89 loc) · 5.95 KB

USER_INSTRUCTIONS.md

File metadata and controls

100 lines (89 loc) · 5.95 KB
Raw

Metly – User Action Checklist

Overview

This document lists all the actions you need to take to get the Metly application fully functional with:

  • AI Proxy (Cloudflare Worker → OpenRouter)
  • Real‑time price feed (Cloudflare Worker)
  • Watchlist & Portfolio persistence (Firebase Firestore)
  • Deployment of the Flutter web app (Vercel)

Each step includes what to do, where, when, and why. Follow the order; later steps depend on earlier ones.

1️⃣ Deploy the AI Proxy Cloudflare Worker

When Where Why
Now (after code changes) Cloudflare Dashboard → Workers & Pages → Create a Worker (e.g., metly-ai-proxy) Provides a secure gateway that adds your METLY_PROXY_TOKEN and forwards requests to OpenRouter.
Paste the content of backend/ai‑proxy/worker.js into the editor.
Settings → Variables → add two environment variables:
• OPENROUTER_API_KEY – your OpenRouter API key. Needed for authentication with OpenRouter.
• METLY_PROXY_TOKEN – a strong random string (you will also store this in the app). Used by the Flutter app to prove it is authorised.
Click Save and Deploy. The worker is now live.
Copy the Worker URL (e.g., https://metly-ai-proxy.<subdomain>.workers.dev/v1/chat/completions). You will paste this URL into the Metly app settings.

2️⃣ Deploy the Price Feed Cloudflare Worker

When Where Why
Immediately after step 1 Cloudflare Dashboard → Create another Worker (e.g., metly-price-feed). Supplies gold & silver prices to the app (currently simulated).
Paste backend/price‑worker/worker.js.
No env vars are required for the simulated version. (If you later switch to a real API, add the needed keys here.)
Deploy and copy the Worker URL (e.g., https://metly-price-feed.<subdomain>.workers.dev).

3️⃣ Configure the Flutter App (Metly Web)

When Where Why
After you have both Worker URLs Open the deployed Metly web app (the Vercel URL you already have). The app reads these settings at runtime.
Go to Settings → AI & Proxy.
• Proxy URL → paste the AI Proxy Worker URL from step 1.
• Proxy Token → paste the METLY_PROXY_TOKEN you created.
Scroll to Data Source.
• Price Worker URL → paste the Price Feed Worker URL from step 2.
Save the settings.
The app will now use the proxy for AI calls and fetch live (simulated) metal prices from the price worker.

4️⃣ Set Up Firebase Firestore (Watchlist & Portfolio)

When Where Why
Before deploying the web app (or immediately after step 3) Firebase Console → your project (metly‑62916). Stores user‑specific watchlist and portfolio data.
Enable Firestore → Create database → choose Start in test mode. Allows reads/writes without authentication for now (development).
(Optional) Later replace test‑mode rules with the secure rules you will write in firestore.rules.
The code in lib/services/firestore_service.dart reads/writes under users/<test-user-id>/watchlist and users/<test-user-id>/portfolio. Hard‑coded test-user-id is used until you add Firebase Auth.

5️⃣ Deploy the Updated Flutter Web App to Vercel

When Where Why
After steps 1‑4 are complete Local machine (project root). Publishes the UI with the new functionality.
Run the following commands (PowerShell): 
flutter build web --release
npx -y vercel@latest deploy --prod

| The first builds the static web assets into build/web. The second uploads them to Vercel. | | | When Vercel asks for the output directory, type build/web. | | | After deployment, you will receive a new Vercel URL. |

6️⃣ Verify the Whole System

When What to Check
Immediately after deployment Open the Vercel URL in a browser.
• Prices appear and update when you tap Refresh Prices.
• Click Ask AI – you should receive a short insight (no errors).
• The Watchlist and Portfolio sections show “No … items” (empty is expected).
• Open the browser console – there should be no CORS or network errors.
If anything fails, revisit the corresponding step and ensure URLs and tokens are correct.

7️⃣ Future Enhancements (Optional, not required now)

  • Add Firebase Authentication and replace the hard‑coded test-user-id with the authenticated UID.
  • Swap the simulated price worker for a real metal‑price API (e.g., MetalPriceAPI) – just modify backend/price‑worker/worker.js to call the external API and return the JSON shape.
  • Harden Firestore security rules (move out of test mode) – edit firestore.rules to restrict reads/writes to the authenticated user.
  • Create UI for adding/removing watchlist items and managing portfolio holdings.
  • Implement push notifications for “Buy” signals (e.g., using Firebase Cloud Messaging).

Quick Reference Table

Step Action Location Key Value
1 Deploy AI Proxy Worker Cloudflare → Workers METLY_PROXY_TOKEN, OPENROUTER_API_KEY
2 Deploy Price Feed Worker Cloudflare → Workers (none required for simulated version)
3 Set URLs in Metly Settings Metly web app → Settings Proxy URL, Proxy Token, Price Worker URL
4 Enable Firestore (test mode) Firebase Console → Firestore –
5 Deploy Flutter web Local terminal → Vercel build/web directory
6 Verify Browser (Vercel URL) –

That’s everything you need to do. Follow the steps in order, and you’ll have a fully‑functional Metly app with AI, live price data, and persistent watchlist/portfolio.