Psyco-Assistant

# Promblem
- The presence of problems, questions
- Skepticism
- Lack of time
- The desire to try cheap or for free

# Solution
The idea to make a psyco-assistant, that analysis the user's problems and help to find out an answers, solve the problems fast in digital space.

## Architecture

### Chat
Main layer that interacts with user, analysis his thoughts, problems, requests, ideas and collects them into context file.
Than the context file sends to the Router layer to save the context of chat layer clear as possible and devide the analysing and decision logic.

### Router
Analysis the file with collected context and make the decision of what kind of App should try the user.

### Mini-Apps
Apps that contains some business-logic and integrated LLM-model that analysis the answers and helps to expand, explore, or deepen beliefs.

#### Business-logic
Some business-logic inside of apps.

#### AI-agent
LLM that analysis the answers, collect the context into the file and send it to chat layer.

## Tech-stack
- FastAPI
- Next.JS
- Redis
- PostgreSQL
- LLM Provider (OpenAI/Google/Meta etc)

- Some local models for testing (LMStudio/Ollama)

No files, tests, or concrete code entry points are identified; the issue describes a proposed architecture using FastAPI, Next.js, Redis, PostgreSQL, and local models such as Ollama. Start by reviewing the architecture and defining a specific MVP, implementation entry point, and acceptance criteria before development begins.

Written by the indexing model from the issue text.