SAAS · SUBMISSION PIPELINE
THE PAPERWORK BEHIND THE papers
- Project
- ResearchGo.ai
- Role
- Architecture & engineering
- Stack
- Next.js · Supabase · Python + FastAPI · Stripe
- Status
- STATUS: IN DEVELOPMENT — FOUNDATION PHASE
Problem
Academic submission pipelines run on email threads and spreadsheets — manuscripts, authors, versions and journal requirements drift out of sync the moment more than one person touches them.
Solution
A Next.js app on Supabase with EU data residency, and a Python worker service behind a signed, replay-protected job contract. The foundation phase shipped what can't be retrofitted — auth including ORCID, abuse-resistant trials, billing, a tested guard on every AI path. The parsing pipeline (GROBID on arrival) is designed and scheduled, not improvised.
Interaction
Security went in before features: every AI call passes a layered guard, every worker job carries a signature and a timestamp.
Outcome
Foundation phase built and tested — migrations, CI and database invariants included. The pipeline phase is next; it launches when it's real, not before.
DECISION: FOUNDATIONS BEFORE FEATURES — AUTH, BILLING AND THE SECURITY SPINE SHIP FIRST.