ModelOD

Serving ML Models On Demand for HPC Users — a Next.js frontend backed by a Go API. Inspired by model-serving catalogs like Tamarind Bio.

Features

• Browse a catalog of servable models
• Search by name, description, author, or modality
• Filter by function (category) and modality
• Model detail pages with an inline "Run inference" form
• Go REST API for models, filters, and job submission

Project structure

backend/    Go REST API (net/http, no external deps)
frontend/   Next.js 14 App Router + TypeScript

Prerequisites

• Go 1.22+
• Node.js 18+

Run the backend

cd backend
go run .
# listens on http://localhost:8080

API

Method	Path	Description
GET	/api/health	Health check
GET	/api/models	List models (q, category, modality query params)
GET	/api/models/{slug}	Get a single model
GET	/api/filters	Distinct categories and modalities
POST	/api/jobs	Submit an inference job
GET	/api/jobs/{id}	Get a job by id

Run the frontend

cd frontend
npm install
npm run dev
# open http://localhost:3000

The frontend proxies /api/* to the backend via next.config.js (BACKEND_URL, default http://localhost:8080).

Notes

Job submission currently returns a simulated result. Replace the logic in backend/main.go (handleJobs) with a real inference backend or queue to serve production models.