ANT

Amasa Nagbe Toe

A.N.T.
Back to Projects

An AI-integrated monorepo for biomedical research workflows

Biomedical AI Platform

A Next.js 14 monorepo pairing Supabase authentication with AI backends, built around a terminal interface, tool handlers, and BullMQ-driven background workers for biomedical research workflows.

Status

In progress — local setup and architectural analysis underway

Tech Stack

Next.js 14 (Monorepo)TypeScriptSupabase AuthBullMQAI Backends

The Story

This platform explores what an AI-native research tool looks like when it's built like real infrastructure rather than a chat wrapper: a terminal-style interface backed by structured tool handlers, background job processing through BullMQ worker queues, and Supabase handling both authentication and data. The current phase is deep codebase analysis — understanding how the terminal, tool handlers, and worker queues interact before extending them.

Major Features

  • Terminal-style interaction model
  • Structured AI tool handlers
  • BullMQ background worker queues for long-running jobs
  • Supabase-authenticated multi-user access

Development Process

  1. 01 Local environment setup
  2. 02 Deep codebase and architecture analysis
  3. 03 Feature-by-feature extension

Challenges

  • Understanding and safely extending a large, pre-existing monorepo
  • Coordinating asynchronous AI tool calls with reliable background job processing

Solutions

  • Methodical, feature-by-feature codebase analysis before making changes
  • Mapping the relationship between the terminal, tool handlers, and worker queues before extending any one piece

Future Roadmap

Extended tool handler libraryProduction deploymentHealthcare-specific workflow modules