A playground for exploring different languages and frameworks by implementing the same application spec in each one.
The application is a barebones Reddit-like API: users can create Posts, leave Comments on those Posts, reply to Comments, and cast upvotes or downvotes on both Posts and Comments.
Every backend implementation follows the same specification and exposes the same REST API, making each one a drop-in replacement for any other.
A single shared Postgres instance handles data storage across all services.
The schema and stored functions are maintained by a local Rust crate, mlad-db/.
The local development environment is orchestrated with Tilt and Docker Compose, providing live reloading and easy switching between backend services.
backends/ # One sub-directory per backend implementation
frontends/ # One sub-directory per frontend implementation
mlad-db/ # Rust crate used to manage database schema
mlad-manage/ # Rust crate used for general management commands across the project
integration/ # Playwright integration tests
data/ # Supporting data files
| Directory | Language | Framework |
|---|---|---|
fastapi |
Python | FastAPI |
| Directory | Language | Framework |
|---|---|---|
nextjs |
TypeScript | NextJs |
- Implement the same spec faithfully in each language and framework
- Highlight idiomatic differences in how each ecosystem approaches the same problem
- Keep each implementation self-contained and independently runnable
- Learn things
- Have fun
- ???
- Profit?
You'll want the following languages installed on your local machine if you want to develop and test code in your local environment:
Otherwise, you are free to run processes via their Docker entrypoints.
Dev container definitions coming soon
To get started, if using Homebrew, you can install the tools listed in the Brewfile using the following command:
brew bundle installOtherwise, you can manually install each of the following:
- just - command runner used to manage pretty much everything in the project
- pre-commit - a framework for tools used to check code quality
- uv - Python package manager and workspace tool
You will also need:
- Docker, particularly Docker Compose
- recommendation: install Docker Desktop to install Docker, Compose, and other tools together
- Tilt to orchestrate the full application on your local machine
- recommendation: follow their installation instructions, including the setup for Kubernetes in Docker.
Make a copy of the different .env.example files in each service:
- Root
.env.example - Backends:
- FastAPI:
backends/fastapi/.env.example
- FastAPI:
Rename these copies to .env to store local environment variables.
Note
The defaults in each .env.example file should be sufficient in most cases,
but I recommend at least changing the values of
PGADMIN_EMAIL and PGADMIN_PASSWORD
in the root env file.
When you start the pgadmin interface,
you will need to login with these credentials.
Of course, you can just run whatever other application you want
to access the postgres database
exposed on port 5432.
To start all services at once, simply run:
tilt upTilt runs everything configured in our Tiltfile, which essentially runs on top of Docker Compose (using compose.yaml). Resources are defined in the Docker Compose spec, while Tilt provides some means for extra local resources, such as migrating the shared database schema automatically and loading fixture data.
With tilt up running in a dedicated console, hit the Space key to open its web UI
and view the logs for each service.
Different services also expose their ports independently, and you can open them from the links provided in the UI.
A pgadmin instance starts alongside the other services in this app, which you can reach at http://localhost:5050 (or follow its link from the Tilt web UI).
The environment variables PGADMIN_EMAIL and PGADMIN_PASSWORD
(see .env.example)
define the login credentials for this interface.
Please see TESTING.md for details.
Key design decisions are documented in docs/adrs/.
Code in this repository was developed with AI assistance, however all code here is (eventually) reviewed by a human and continually refactored to improve quality, clarity, coherence, and idiomatic patterns in their respective frameworks.
This project is not intended to be deployed into any production environment. Rather, it serves for education (of its author!) and demonstration. If it was intended for production use, the author of this work would have much more strict guidelines here.
Any contributions to this repository should adhere to the same standard set by the above disclosure. Human contributors must always be directly accountable for the work committed.
Contributions suspected of being created and submitted solely by an AI agent will be summarily closed with a reminder about this standard.