A self-hosted FastAPI backend built as a general-purpose service layer for web apps, automation, real-time interaction, authentication, monitoring, and experimental AI-connected systems.
FastAPI, async endpoints, WebSockets, structured models and schemas.
Docker, MicroK8s, MariaDB, ingress, Nginx Proxy Manager and deployment manifests.
JWT authentication, role-based access control, TLS/SSL and API-oriented access patterns.
Designed to expose backend capabilities to AI clients and experimental self-hosted logic systems.
This project reached a live operational state. A custom GPT called App Architect and Coder was configured to make API calls against the backend, demonstrating a working path from natural-language interaction to services hosted on infrastructure I controlled.
/jc-web-api ├── app │ ├── api │ │ ├── analytics.py │ │ ├── db_operations.py │ │ ├── deployments.py │ │ ├── discord.py │ │ ├── health.py │ │ ├── notifications.py │ │ ├── payments.py │ │ ├── users.py │ │ └── ws/ │ ├── core │ │ ├── auth.py │ │ ├── config.py │ │ ├── database.py │ │ ├── logging.py │ │ └── sentient_ai.py │ ├── models/ │ ├── schemas/ │ └── main.py ├── database/migrations ├── domains/ ├── k8s/ ├── logs/ ├── scripts/ └── tests/
The project gradually expanded beyond a single API. It began to look more like a general AI-controlled application platform: a backend that could expose services, state, automation, monitoring, and application behavior to multiple interfaces. Later work such as sentient-proxy explicitly referenced this backend as part of a larger self-hosted system.