A personal project where I utilise the main OpenAI and Deepseek API premium models. Calling the API is way less pricey than paying for premium subscriptions!

Since this is a paid API, if you would like access, please email me at haroldtm55@gmail.com, and I'll grant you free access.

This is a Next.js 15.5.15 App Router project bootstrapped with create-next-app.

Example using gpt-4o text generation model

Provides text outputs in response to their inputs.

Example using gpt-4o vision model

Upload an image and it will answer any questions about it.

Example using dall-e-3 image generation model

Create images from scratch based on a text prompt.

Example using whisper-1 speech-to-text model

Transcribe audio into whatever language the audio is in.

Example use the Assistant API to build an Associate Cloud Engineer Assistant

Upload any screenshot containing an ACE exam question and it will give you the answer.

Getting Started in Local

Note: Since this is a paid API, you need an access key. Feel free to contact me and I'll provide you with one.

Install dependencies and run development server:

pnpm install --frozen-lockfile
pnpm dev

Open http://localhost:3001 with your browser to see the result.

Docker

Build image locally

docker build -t openai-models .

Run image locally (with runtime env):

docker run --env-file .env.local -p 3001:3001 openai-models

Production deploy on VM (recommended)

Use Docker Compose + runtime env file. Do not bake secrets into image.

One-time VM setup:

mkdir -p ~/openai-models-deploy
cd ~/openai-models-deploy

Copy deployment files from this repo to VM deploy directory:

scp -i openai-models-oracle.key docker-compose.prod.yml ubuntu@<VM_IP>:~/openai-models-deploy/
scp -i openai-models-oracle.key scripts/deploy-vm.sh ubuntu@<VM_IP>:~/openai-models-deploy/
scp -i openai-models-oracle.key .env.openai-models.example ubuntu@<VM_IP>:~/openai-models-deploy/.env.openai-models

On VM, edit env values:

cd ~/openai-models-deploy
nano .env.openai-models

Required values:

• OPENAI_API_KEY
• AUTH0_SECRET
• AUTH0_BASE_URL (must be public app URL, e.g. http://152.67.112.83:3001)
• AUTH0_ISSUER_BASE_URL (must be valid URL, e.g. https://your-tenant-region.auth0.com)
• AUTH0_CLIENT_ID
• AUTH0_CLIENT_SECRET
• DEEPSEEK_BASE_URL
• DEEPSEEK_API_KEY

Important:

• Keep values unquoted (no wrapping '...' or "...").
• Keep .env.openai-models on VM only.

Run deploy on VM:

cd ~/openai-models-deploy
chmod +x deploy-vm.sh
./deploy-vm.sh

Normal update flow (next time)

1. Push code to GitHub.
2. Wait for GitHub Action to build and push haroldtm55/openai-models:latest.
3. SSH to VM and run:

cd ~/openai-models-deploy
./deploy-vm.sh

Quick checks on VM:

docker compose -f docker-compose.prod.yml ps
curl -I http://localhost:3001/api/auth/login