Aisé

adding reminders shouldn't be the hard part.

Comment

Aisé

creating reminders shouldn't be the hard part.

Python Django SQLite

What is Aisé?

Prototype Demo

Inspired by the ease of use of TypeForms, and from a hatred of the disgusting Apple Calendar Event forms (sorry Apple, y'all been falling off), Aisé aims to be a useful way for users to quickly create reminders, so they can get to work, rather than waste time crafting the perfect reminder.

It utilizes the Django REST framework and a frontend designed by the ground-up (by a human), in Vanilla HTML and CSS, for creating, listing, and parsing reminders. It includes optional local LLM integration (via Ollama) to parse natural-language reminder text into structured reminder data (title, date, time, location).

Features

  • Reminders CRUD: REST endpoints to create, read, update and delete reminders using Django REST Framework.
  • AI parsing: /api/ai/parse/ endpoint uses a local Ollama model to extract structured fields from free-text reminders. Falls back to Python heuristics when Ollama is not available.
  • Title polishing: /api/ai/polish/ endpoint to clean up reminder titles (uses Ollama when available, otherwise local heuristics).
  • Simple frontend: Single-page app served at the root to demo the UI (templates/Aisé/app.html).
  • User-aware storage: Reminders may be associated with authenticated users; anonymous/demo usage supported.

Tech stack

  • Python 3.11+
  • Django + Django REST Framework
  • SQLite (default development DB)
  • Optional: Ollama + Qwen 3.5 (for AI parsing)

Project layout

  • manage.py — Django management entrypoint
  • config/ — Django project settings and URL config
  • Aisé/ — main app: models, serializers, views, templates, static files
  • db.sqlite3 — default SQLite database (development)

API endpoints

  • GET /api/reminders/ — list reminders (DRF)
  • POST /api/reminders/ — create reminder
  • POST /api/ai/parse/ — parse free-text reminder into JSON
  • POST /api/ai/polish/ — polish a reminder title

These are registered in config/urls.py and implemented in Aisé/views.py.

Running locally (development)

  1. Create a virtual environment and install dependencies:
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
  1. Apply migrations and run the dev server:
python manage.py migrate
python manage.py runserver
  1. Open the frontend in your browser:
http://127.0.0.1:8000/
  1. (Optional) Use Ollama for better parsing:
  • Install Ollama and pull an appropriate model (e.g. qwen3.5).
  • Ensure the ollama Python package is installed in your environment.

If Ollama is not installed or the model is unavailable, the AI endpoints fall back to built-in Python parsing heuristics.

Configuration & secrets

This repo currently uses a minimal local setup. If you add external services, keep credentials out of source control and load them via environment variables or a .env/secrets.env file.

Limitations

  • Intended as a lightweight demo / starter app — not hardened for production use.
  • No production authentication, rate-limiting, or strict validation beyond basic serializers.
  • Ollama integration is optional and requires a compatible local environment.

Next steps and ideas

  • Add user registration / authentication flows (JWT or social login).
  • Add tests for AI fallbacks and serializers.
  • Deploy with Docker and add a production-grade DB and static file serving.
  • Add a small CI check that runs linters and basic tests.

NOTES

  • UI was completely me (Vanilla HTML and CSS)
  • Django REST framework was also me
  • Claude was used solely for the javascript, which will be done by me (it's trash) once I get home
  • yes a lot of this README was generated as well 😭
Share this project:

Updates