A web app that helps you pick compatible PC parts for gaming — choose a budget, pick the games you want to play, and it'll suggest the best builds for the money. Also handles upgrades for existing PCs.
Built as an A-Level Computer Science NEA project (OCR), originally under the name BuildLogicPCS.
- New Build — tell it your budget, select a game and quality preset (Ultra/High/Medium/Low), and it generates up to 6 complete PC builds ranked by performance. Each build shows component breakdown, total price, logic score (its performance metric), and any CPU/GPU bottleneck.
- Upgrade — if you already have a PC, pick your existing components and a budget. It'll find the best upgrades (CPU, GPU, RAM, or combo) that fit the budget and give the biggest performance lift.
- Compare — pick two builds side-by-side and see component differences, price, scores, and bottlenecks.
- Case Designer — after picking a build, choose a case that fits the hardware, with a live preview and price update.
- Purchase Links — every component gets a static Google Shopping search link so you can check current prices.
Everything is checked for compatibility automatically — CPU socket matching, RAM generation, PSU wattage, PCIe version. No manuals needed.
| Thing | What |
|---|---|
| Backend | Python + Flask (3.1) |
| Database | SQLite (embedded, ~9MB of component data) |
| Frontend | HTML + CSS + vanilla JS |
| Scoring | Weighted logic score (CPU 40%, GPU 50%, RAM 10%) + game-requirement matching |
The component database ships pre-loaded with 98 CPUs, 98 GPUs, 50 RAM sticks, 20 motherboards, 20 PSUs, 29 cases, and 15 games with 4 quality presets each. All performance scores are stored from real benchmarks.
| Homepage | New Build Config | Results | Upgrade |
|---|---|---|---|
 |
 |
 |
 |
# Clone
git clone https://github.com/TheAhmadZeb/pc-builder-optimiser.git
cd pc-builder-optimiser
# Install Flask
pip install flask
# Run
python3 app.pyOpen http://localhost:5000. The database builds itself on first run from the SQL files in sql/.
pc-builder-optimiser/
├── app.py # Flask routes
├── backend/
│ ├── db.py # SQLite connection & queries
│ ├── compatibility.py # CPU/RAM/PSU/PCIe compatibility checks
│ ├── scoring.py # Logic score, bottleneck, game performance
│ ├── optimisation.py # Build generator + upgrade recommender
│ └── comparison.py # Side-by-side build comparison
├── sql/
│ ├── schema.sql # Database schema
│ └── seed.sql # Pre-loaded component & game data
├── static/
│ ├── css/style.css
│ ├── js/app.js
│ ├── fonts/ # Custom fonts
│ └── assets/ # Images, game posters, case previews
├── templates/ # Jinja2 templates
└── screenshots/ # App screenshots
- Compatibility checking — socket matching, RAM type, PSU wattage, PCIe version
- Game performance estimation — checks CPU/GPU scores against stored game requirements with dynamic weighting based on whether each game is CPU-heavy or GPU-heavy
- Logic score — weighted combination of CPU (40%), GPU (50%), RAM (10%) performance scores with a geometric mean penalty for unbalanced builds
- Build optimisation — iterates through component combinations within budget, picks the top 6 by logic score
- Bottleneck detection — compares CPU vs GPU performance to identify the limiting component and estimate severity
- Upgrade recommendation — evaluates all CPU, GPU, RAM upgrade paths within budget, returns the best options
- No live pricing — prices are static snapshots. Useful for comparison, not up-to-the-minute shopping.
- No real FPS benchmarks — bottleneck percentages and performance scores are estimates based on stored data, not real-time game tests.
- No user accounts or saved builds — everything lives in the browser session.
These were deliberate trade-offs to keep the project self-contained and testable.
MIT — use it, break it, improve it.