Option A — Phase 0: Prerequisites (Detailed Plan)

This document is the detailed plan for Phase 0 of AppiFire AI Chat (Option A). Complete these before starting Phase 1. High-level overview: Option-A-Tech-Stack-and-Tasks.md.

---

Phase 0 objective

Have all accounts, tools, and infrastructure ready so you can create the Shopify app and connect it to Postgres and OpenRouter in Phase 1.

---

1. Create Shopify Partner account and development store

1.1 Create Shopify Partner (developer) account

Step	Action
1	Go to partners.shopify.com.
2	Click Join now.
3	Sign in with email or Google.
4	Choose: Account type: Developer. Company name: your product or company name. Purpose: Build apps for the Shopify App Store.

Outcome: You land in the Shopify Partner Dashboard. Here you will create apps, development stores, and manage API keys.

1.2 Create a development store

Step	Action
1	In Partner Dashboard, go to Stores.
2	Click Add store.
3	Select Development store.
4	Store details: Store name (e.g. appifire-ai-chat-test), Password protection: enabled (default).
5	Click Create store.

Outcome: A real Shopify store, free, with full admin. It can install unpublished apps — use it for all Phase 1 testing.

Checklist:

• Partner account created and logged in.
• At least one development store created; note the store URL (e.g. your-store.myshopify.com).

---

2. Install Node.js 18+ and Shopify CLI; run shopify login

2.1 Install Node.js

Step	Action
1	Install Node.js 18 or later (nodejs.org or via nvm/homebrew).
2	Verify in terminal: node -v and npm -v.

2.2 Install Shopify CLI and App package

Step	Action
1	Run: npm install -g @shopify/cli @shopify/app
2	Verify: shopify version

2.3 Log in to Shopify via CLI

Step	Action
1	Run: shopify login
2	In the browser that opens, sign in with the same email as your Partner account.
3	Confirm in terminal that login succeeded.

Outcome: CLI can create apps, deploy extensions, and run the local dev server against your Partner account.

Checklist:

• Node.js 18+ installed (node -v).
• Shopify CLI and app package installed (shopify version).
• shopify login completed successfully.

---

3. Sign up for OpenRouter.ai; get API key; add credits

3.1 Create account and get API key

Step	Action
1	Go to openrouter.ai.
2	Sign up / log in.
3	Open Keys (or API keys section) and create a new API key.
4	Copy and store the key securely (you will use it as OPENROUTER_API_KEY in the backend).

3.2 Add credits

Step	Action
1	In OpenRouter dashboard, go to Billing or Credits.
2	Add credits (amount depends on your testing plan).

Why: All AI usage (embeddings + chat) for Option A goes through OpenRouter; one bill and one key for both.

Checklist:

• OpenRouter account created.
• API key created and saved securely.
• Credits added to the account.

---

4. Choose hosting for backend + Postgres; provision Postgres with pgvector

4.1 Choose a hosting strategy

You need:

• A PostgreSQL database with the pgvector extension (for vectors).
• A place to run your backend app (Phase 1). It can be the same provider or different.

Common options:

Option	Postgres + pgvector	Backend host	Notes
Railway	Railway Postgres (pgvector supported)	Same project	Simple; one dashboard.
Render	Render Postgres (add pgvector)	Web Service (Node)	Free tier for DB; paid for always-on.
Supabase	Supabase (Postgres + pgvector)	Separate (e.g. Fly.io, Render)	Good free tier; DB and app decoupled.
Neon	Neon (Postgres + pgvector)	Separate (e.g. Fly.io, Render)	Serverless Postgres; good for dev.
DigitalOcean	Managed DB (add pgvector)	Droplet or App Platform	More control; you manage.
Fly.io	Fly Postgres (pgvector)	Fly app	Single provider for both if desired.

Recommendation for Phase 0: Pick one provider and create only the database now. Create the backend app in Phase 1 and connect it to this DB.

4.2 Provision Postgres and enable pgvector

Step	Action
1	In your chosen provider, create a new PostgreSQL database.
2	Note the connection string (e.g. postgresql://user:pass@host:5432/dbname). You will use it as DATABASE_URL.
3	Enable pgvector: run CREATE EXTENSION IF NOT EXISTS vector; in the DB (via provider SQL console or psql). For schema-qualified vector types use CREATE EXTENSION IF NOT EXISTS vector WITH SCHEMA public;.

Docker / Dokploy / self-hosted: Use a Postgres image that includes pgvector (e.g. pgvector/pgvector:pg15 or pgvector/pgvector:pg16). Standard postgres:15 does not include it; the extension will be missing and you will get “extension vector is not available”.

Verification: Connect with psql or a GUI and run:

CREATE EXTENSION IF NOT EXISTS vector WITH SCHEMA public;
SELECT * FROM pg_extension WHERE extname = 'vector';

You should see one row for vector.

4.3 (Optional) Create an empty database or schema

Some providers create an empty DB by default. If not, create a database (e.g. shopify_ai_chatbot) for the app. Do not create tables yet — Phase 1 will add migrations for the full schema.

Checklist:

• Hosting provider chosen for Postgres (and later backend).
• Postgres instance created; connection string saved as DATABASE_URL.
• pgvector extension enabled (CREATE EXTENSION vector).
• (Optional) Dedicated database created for the app.

---

Phase 0 completion checklist

Before starting Phase 1, confirm:

• Shopify: Partner account + at least one development store.
• Local: Node.js 18+, Shopify CLI installed, shopify login done.
• OpenRouter: Account, API key saved, credits added.
• Database: Postgres provisioned, pgvector enabled, DATABASE_URL known.

Next: Option-A-Phase-1.md — Shopify app skeleton and database migrations.