momo2

Local-first AI digital assistant with:

• Python backend
• web frontend
• one user-facing chat agent
• internal specialized sub-agents
• LM Studio as the default model provider
• OpenRouter as fallback
• Docker-based local and deployment workflow
• Forgejo CI/CD

This repository is currently scaffolded for planning-first implementation. The initial goal is to give human and AI contributors a clear architecture, task breakdown, testing strategy, and verification workflow before feature work begins.

Planned Stack

• Backend: FastAPI, SQLAlchemy, Pydantic, Redis, Postgres
• Frontend: React + TypeScript + Vite
• E2E testing: Playwright
• Local orchestration: Docker Compose
• CI/CD: Forgejo Actions / pipeline

Repository Layout

apps/
  api/        Python backend
  web/        React frontend
docs/         Architecture, requirements, and runbooks
infra/        Docker and Forgejo support files
tests/        End-to-end and shared test fixtures
AGENTS.md     Top-level instructions for coding agents

Current Status

The repository contains:

• planning documents in docs/
• coding-agent instructions in AGENTS.md
• placeholder application skeletons in apps/
• Docker and CI/CD stubs for future implementation

Next Steps

1. Finalize architecture and task ordering in the docs.
2. Implement the backend service skeleton.
3. Implement the frontend chat shell.
4. Add provider adapters for LM Studio and OpenRouter.
5. Build agent orchestration and verification flows.