A practical laptop guide for Indian B.Tech students. Match a machine to your branch, projects and budget—with the reasons behind every score.
Set the requirements that matter to you. Scores are workload-fit estimates—not synthetic benchmark percentages.
One codebase, two hosting modes: Cloudflare Pages can use Gemini through a private server-side function; GitHub Pages and offline visits automatically fall back to the bundled local rules engine.
Auto mode tries the same-origin api/gemini endpoint for up to 28 seconds; the server-side request times out earlier. If the endpoint is missing, times out, hits quota, or returns an error, this browser automatically uses the bundled rules engine. Use Check cloud endpoint to confirm the Function is deployed and the secret is present (this check does not validate the API key). Optional Ollama runs a model on your own computer.
Privacy note: Offline mode processes your question in this browser. In Auto mode, if the same-origin Cloudflare endpoint is configured, your question plus limited workload/budget and laptop-summary data is sent to Google Gemini through that endpoint. The API key stays server-side; Google API free-tier terms and quotas apply. Use Offline mode to avoid sending the question to a cloud AI service. Ollama mode contacts only http://localhost:11434.
Add up to three laptops to compare RAM, graphics, upgradeability and workload-fit notes side by side.
Use “Compare” on a laptop card to add a model.
The engine matches known specs against a transparent workload profile. It does not pretend to measure real-world speed from a product name.
Different weights for coding, AI/ML, CAD, gaming, portability and everyday study.
Cards explain when RAM, VRAM, upgradeability or unknown specifications could be a limitation.
Every known price is a dated snapshot. Retailer links help you confirm the exact SKU before buying.
Important: CPU/GPU benchmark values are reference snapshots, not laptop-level guarantees. “Software headroom” is a weighted heuristic, not a promise of how many years a laptop will last. Exact model, power limits, cooling, and configuration matter.
Update the catalogue locally in VS Code. Import a JSON file to preview changes, then export or replace the file in your repository yourself.
Keep laptop specs and dated prices in separate files. Prices are blank until a real listing and observation date are checked. Benchmarks and model specs include their source links.
Import a JSON catalogue to test it in this browser. This preview does not modify your repository files automatically.
No custom file loaded.
Use Amazon India, Flipkart and Google Shopping search links to check current offers. Benchmark numbers are dated external measurements; laptop power limits and cooling can change real performance.
PassMark CPU benchmark database ↗PassMark GPU benchmark database ↗Local JSON stays the primary catalogue. Paste a publicly published Google Sheets CSV URL to merge non-empty cells by laptop ID. This requires internet only when you click sync; no Google API key is required.
No sheet source configured.
Download CSV header template ↓Publishing makes that sheet tab publicly readable. Never place secrets or private student data in it. Changes are in-memory until you export the catalogue and replace the JSON file in VS Code.