echo-mind

A portfolio RAG chatbot. Visitors ask questions about my projects, blog posts, and experience; the bot answers using retrieval-augmented generation over my markdown content.

Deployed on AWS: API Gateway → Lambda (container) → FastAPI → (Phase 3) Claude API.

Project status

Phase	Description	Status
1	Infrastructure scaffolding	🚧 Deploy to complete
2	Document ingestion & embeddings	⏳ Not started
3	Retrieval & RAG chain	⏳ Not started
4	Chat memory & frontend integration	⏳ Not started

See docs/02-phase-plan.md for the full learning roadmap.

Quick start

Prerequisites

• AWS CLI configured (aws sts get-caller-identity should work)
• Terraform >= 1.6
• Docker (with BuildKit)
• Python 3.12

First-time deploy (Phase 1)

# 1. Bootstrap remote state + auto-generate backend.tf (one-time per account)
./scripts/initialize_aws.sh

# 2. Configure variables
cd infra/envs/dev
cp terraform.tfvars.example terraform.tfvars
# edit terraform.tfvars — set your name in common_tags

# 3. Init
terraform init
cd ../../..

# 4. Deploy (builds image, pushes to ECR, applies Terraform)
./scripts/deploy.sh

Local development

./scripts/local-dev.sh
# Then: curl http://localhost:8000/health

Run tests

cd backend
pip install -r requirements-dev.txt
pytest tests/

Structure

echo-mind/
├── infra/          Terraform modules and environment compositions
├── backend/        FastAPI app + Lambda Dockerfile
├── frontend/       Chat UI (HTML/CSS)
├── scripts/        deploy, local-dev, bootstrap helpers
└── docs/           Architecture diagram and phase plan

Architecture

See docs/01-architecture.md.