Building Ascend Spatial Labs — The Technology Behind the Platform
Jimmy MatewereAscend Spatial Labs was founded by Nathaniel Phiri to close the infrastructure data gap across Southern Africa. I joined as CTO and built the platform from scratch. No clients yet, which means every architectural decision was made on judgment rather than feedback.
The stack is Next.js, Sanity, Clerk, Resend, and Mapbox. Clerk was the least obvious choice. Auth is the kind of problem that looks simple until you're deep in it. NextAuth requires you to own the UI and wire everything manually. Rolling your own means owning the security surface entirely, which at an early stage is mostly a liability. Clerk ships prebuilt components, handles token rotation and session management out of the box, and the route protection pattern using clerkMiddleware with createRouteMatcher is clean to read and reason about. Deny-list by default, explicit public routes, safer posture than allow-listing. It's a managed service with real pricing implications at scale, but for now it's the right call.
Sanity is doing double duty as CMS and database, which is unorthodox. The concern isn't the current state, it's what happens when real project datasets start coming in. For now it works, and there's a useful side effect: client data stays backed up somewhere we control rather than depending on what a client has locally.
The Clerk-to-Sanity sync runs through a webhook. When an account is deleted it doesn't get removed from Sanity. It shifts from Active to Archived. Pending and Suspended states exist too. It's held up across every test account I've run through it. Real operational load is still an open question.
Some features on the portal are ahead of where the company actually is. The project-level report generator makes sense, it's tied to a deliverable. The second report, the one aggregating data across all projects with growth charts, is solving a problem I imagined. It'll either get replaced or justify itself once real clients are using the system.
Domain purchase and newsletter are next. Ascend could go in directions we haven't fully mapped, and being able to communicate that directly to people paying attention is worth building early.
Continue the Conversation
I welcome peer perspectives and questions regarding any of the topics discussed.

