OmniAI — LLM Comparison Terminal (Prototype)

OmniAI is a terminal-inspired dashboard for comparing large language models across providers. It brings pricing, context length, speed, benchmark scores, modalities, and best-fit tags into one searchable interface.

Selecting a model for a product often means switching between provider pages and manually reconciling different pricing formats, context limits, and capability claims.

Browse a typed dataset covering 16 models from 16 providers.

Search, filter by provider, and sort by price, speed, context size, or rating.

Compare model cards and scan a scrolling price ticker.

Explore routes designed for a leaderboard, AI news aggregation, and pricing snapshots/history.

Built with Next.js 16, React 19, Tailwind CSS 4, Recharts, and shadcn/ui components. The project includes a typed model schema, SWR data fetching, a Neon serverless PostgreSQL schema for pricing snapshots, and a cron-secret-protected snapshot endpoint. It is configured for Cloudflare Pages.

The app was run locally; the main screen rendered without console errors.

Model names, prices, and ratings are sourced from a static curated dataset. The interface's “live” label does not mean the data updates live, and values may be outdated.

PostgreSQL-backed pricing history, external leaderboard calls, and news routes were not exercised during the local run.

No currently working public deployment is linked.