smart-task-api 🤖

An AI-powered task management REST API that automatically prioritizes tasks using a locally-running LLM (Ollama + llama3) — no API keys, no external services needed.

How It Works

1. You create a task with a title and description
2. The API saves it instantly and returns a response with priority: pending
3. In the background, the task is sent to llama3 running locally via Ollama
4. The AI assigns low / medium / high priority plus a one-line reason
5. The task is updated automatically

Tech Stack

Layer	Technology
API Framework	FastAPI 0.115 (async)
Database	PostgreSQL 16 + SQLAlchemy 2
AI Model	Ollama + llama3 (local)
Containerization	Docker + Docker Compose
Testing	pytest + pytest-asyncio
Linting	Ruff
CI	GitHub Actions

Quick Start

git clone https://github.com/JeffiN11/smart-task-api.git
cd smart-task-api
docker compose up --build

API: http://localhost:8000 Docs: http://localhost:8000/docs

API Endpoints

Method	Endpoint	Description
GET	/	Health check
POST	/tasks/	Create task (AI prioritizes in background)
GET	/tasks/	List tasks (filter by status/priority)
GET	/tasks/{id}	Get single task
PATCH	/tasks/{id}	Update task
DELETE	/tasks/{id}	Delete task
POST	/tasks/{id}/reprioritize	Force AI re-prioritization

Example

Create a task:

curl -X POST http://localhost:8000/tasks/ \
  -H "Content-Type: application/json" \
  -d '{"title": "Fix login bug", "description": "Users cannot log in on mobile when 2FA is enabled."}'

Instant response:

{
  "id": 1,
  "title": "Fix login bug",
  "priority": "pending",
  "ai_reasoning": null
}

After AI responds:

{
  "id": 1,
  "title": "Fix login bug",
  "priority": "high",
  "ai_reasoning": "Login failures block all users from accessing the product."
}

Project Structure

Running Tests

pip install -r requirements.txt -r requirements-dev.txt
pytest -v

AI Prioritization Logic

• high — Urgent, blocking others, production issues, deadlines within 24h
• medium — Important but not urgent, should be done this week
• low — Nice-to-have, minor improvements, no immediate impact

If Ollama is unavailable, the API gracefully defaults to medium priority and never fails.