FastAPI
High-performance Python API gateway—OpenAPI docs, async I/O, and typed request/response models for our workflow, uploads, and health endpoints.
Study Coach is a vertical AI platform for learners in Ghana and across the continent—personalized coaching that understands local programmes, document intelligence on your materials, and an architecture designed for offline-capable use where connectivity is uneven.

High-performance Python API gateway—OpenAPI docs, async I/O, and typed request/response models for our workflow, uploads, and health endpoints.
Checkpointed agent graphs with thread-scoped memory. Routes between reasoning and tools until the coach finishes—backed by SQLite for durable sessions.
App Router, React 19, and a typed client UI for assessment, /studio coach, and Clerk-ready auth—calling the same REST API as our legacy static pages.
Agents call tools over MCP (FastMCP)—kept alongside this stack for discoverable prompts and integrations.
How it works

Level, programme, subject focus, and goals — tuned for GES, SHS, WASSCE paths, tertiary, and digital literacy.

A private learning ID and server-side memory so your coach remembers context across sessions.

No fabricated exam aggregates—we cite WAEC, GES, and your teachers for official rules.
Intelligence layer
Personalized study roadmaps that respect how you learn—sequencing topics and adjusting difficulty.
Uses your history and study patterns to surface gaps early and clear next steps.
Pinpoints concepts that need work and proposes focused practice—not generic drills.
This Next.js app uses Clerk for sign-in when configured. The FastAPI backend validates Clerk session tokens. Add your frontend origin to CORS_ORIGINS and CLERK_AUTHORIZED_PARTIES on the API. Learning IDs are private—treat them like passwords.
Educational support only. Verify exams with official sources. Service map