Dr. Anne Hussain Platform

A unified platform for a naturopathic doctor - book, blog, and podcast in one place, with a RAG assistant that answers questions from her published work and cites its sources.

Visit project

Product screens

Unified Platform

One owned home for her practice, book, blog, and podcast.

Unified Platform

AnneAI Grounded Answers

Answers come strictly from her published work, with numbered citations.

AnneAI Grounded Answers

Podcast Network

A custom player and episode feed; new episodes index into the assistant automatically.

Podcast Network

Book Hub

The Period Literacy Handbook in eBook, paperback, and audiobook formats.

Book Hub
Execution Snapshot

The strongest signal first, then the operating context around it.

Lead Signal

RAG + citationsAI Layer across a shipped digital platform / ai build.

Delivery Role

End-to-end solo build as her long-term developer: product strategy, design, Next.js application, podcast ingestion pipeline, and the embeddings and retrieval system behind AnneAI.

Product Context

Dr. Anne is a naturopathic doctor and the author of The Period Literacy Handbook. Her content lived on platforms she did not own, and as her catalog grew, so did the gap between what she had published and what her audience could actually find.

Book + blog + podcast

Content Hub

Network deal + referrals

Outcome

Launch Posture

The stack and feature set were shaped for production use, not just a polished demo.

Next.jsTypeScriptOpenAIRAGStripePrisma

Build Narrative

A clean story from constraint to shipped outcome.

01

Problem

01

Dr. Anne's audience was split across rented platforms - a book on Amazon, a podcast on hosting services, a separate blog - so each channel grew alone and none of it was searchable together. Readers kept asking questions she had already answered somewhere in her published content.

Constraint mapping
02

Build

02

I built annehussain.com as a single owned platform: a book hub, a podcast network with its own player, and a blog, topped with AnneAI - a retrieval-grounded assistant that answers questions strictly from her published corpus and renders inline numbered citations linking back to the source posts and episodes.

System design
03

Outcome

03

A production platform that turned a fragmented content footprint into one owned surface - and a working example of citation-grounded RAG serving a health expert's real audience.

Production outcome

Framing

Defining the product and the operating constraints.

Dr. Anne is a naturopathic doctor and the author of The Period Literacy Handbook. Her content lived on platforms she did not own, and as her catalog grew, so did the gap between what she had published and what her audience could actually find. I centralized the content into one Next.js app, then built the retrieval layer: every post and episode transcript is embedded into Supabase pgvector, and AnneAI composes answers from retrieved passages with numbered citations. Publishing a new episode updates the assistant's knowledge automatically.

Systems Index

Next.js
TypeScript
OpenAI
RAG
Stripe
Prisma

Key features in scope

AnneAI - retrieval-grounded Q&A with inline numbered citations and a source list
Separate Ask and Search modes over the full blog and podcast corpus
Podcast network pages with a custom player and automated episode ingestion
Book hub for The Period Literacy Handbook - eBook, paperback, and audiobook

Role and product posture

Role: End-to-end solo build as her long-term developer: product strategy, design, Next.js application, podcast ingestion pipeline, and the embeddings and retrieval system behind AnneAI.
Category: Digital Platform / AI

Engineering

Building the core system and choosing where to be opinionated.

I built annehussain.com as a single owned platform: a book hub, a podcast network with its own player, and a blog, topped with AnneAI - a retrieval-grounded assistant that answers questions strictly from her published corpus and renders inline numbered citations linking back to the source posts and episodes.

Systems Index

Next.js
TypeScript
Tailwind CSS v4
Framer Motion
Clerk
Next.js API Routes
Prisma
Stripe

Architecture choices

Next.js App Router application with API routes, deployed on Vercel
Supabase pgvector store holding embeddings for every blog post and podcast transcript
Ingestion pipeline that transcribes new podcast episodes with AssemblyAI and indexes them for retrieval automatically
RAG answer flow that retrieves scored passages, generates the response with OpenAI, and renders inline numbered citations plus a source list
Clerk auth and Stripe wired in for member access and paid content

Key decisions

Bounded AnneAI strictly to her published corpus - health content demands grounded answers with visible sources, not open-ended generation
Automated podcast ingestion end to end so a new episode becomes searchable and citable without manual work
Gave the assistant its own page with separate Ask and Search modes, so visitors choose between a composed answer and raw source retrieval

Hardening

Turning the build into something resilient enough to matter.

A production platform that turned a fragmented content footprint into one owned surface - and a working example of citation-grounded RAG serving a health expert's real audience.

Systems Index

Production RAG with citations for a real client audience, not a demo
A long-running client relationship with legible outcomes - a podcast network deal and practitioner referrals
Full ownership across the content platform, the ingestion pipeline, and the AI layer

Results after shipping

Runs in production as her primary web presence, with AnneAI answering visitor questions with inline citations to her blog and podcast
The unified podcast surface contributed to her securing a podcast network deal
The assistant led to referrals from other practitioners who wanted the same treatment of their own content

Constraints

Health content: answers had to stay grounded in her published material, with sources the reader can check
New podcast episodes had to become searchable without any manual indexing step
The platform serves a real audience and had to stay fast, simple, and maintainable by one developer

Lessons

What the build taught me.

01

RAG over a single expert's corpus is one of the highest-leverage AI features a creator platform can ship

02

Citations are the trust feature - in health content, showing the source matters as much as the answer

03

Automated ingestion is what keeps a knowledge system alive; an index that needs manual updates rots