This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle
What I Built
PrioritiAI is a privacy-first, open-source AI task management and scheduling web application that cuts through decision paralysis for students, freelancers, and small business owners.
The Problem It Solves
Traditional todo apps function as static checklists: they collect tasks, but leave users overwhelmed by the hardest question: "What should I work on right now, and why?" When users face 15+ competing responsibilities, they experience decision fatigue and fall into procrastination or default to low-impact tasks. Meanwhile, closed cloud AI tools demand that users stream sensitive client notes, deadlines, and project details to third-party servers.
The Solution
PrioritiAI introduces a personal AI productivity assistant that:
- Identifies "Your Next Best Task": Evaluates tasks across Urgency, Importance, Effort & Quick Wins, Prerequisite Dependencies, and Mental Energy Requirements, spotlighting the single highest-yield item to tackle next with transparent explanations.
- Accepts Natural-Language Task Dumps: Lets users input thoughts naturally ("Finish my DBMS assignment by tomorrow, prepare for Friday's exam, and spend 30 minutes replying to client emails") and extracts actionable task structures with user review before saving.
- Generates an Honest Daily Schedule ("Plan My Day"): Slots tasks into realistic working hours, inserts recovery breaks, and honestly warns users when workloads exceed capacity rather than overpromising.
- Protects Personal Privacy: 100% of data remains stored locally on device with zero cloud telemetry.
Who Would Pay for It?
- Freelancers & Consultants: Juggling multiple client deliverables with strict billable hour constraints who need to protect client confidentiality.
- Students & Researchers: Balancing exams, assignment deadlines, and study blocks who need automated workload pacing.
- Small Business Owners: Prioritizing operations and revenue-generating activities without expensive monthly SaaS subscription bloat.
Code
The complete source code is open-source under the MIT License on GitHub:
song-code-won
/
PrioritiAI
PrioritiAI - Smart AI Task Planner with deterministic multi-factor priority scoring and local open-weight AI (Ollama/Qwen).
PrioritiAI — Smart AI Task Planner
A privacy-first, open-weight AI productivity assistant that cuts through decision paralysis, calculates transparent priority rankings, and schedules realistic daily workflows.
📖 1. Problem & Intended Users
Students, freelancers, software developers, and small business operators frequently encounter task overload. When faced with 10–20 concurrent responsibilities across different contexts (deadlines, client emails, exam preparation, code reviews), humans experience decision fatigue and default to either procrastinating or tackling low-impact tasks first.
Most existing todo apps are mere static lists: they record what needs to be done, but provide zero guidance on what to do next and why.
PrioritiAI bridges this gap:
- Multi-Factor Priority Scoring: Evaluates urgency, importance, effort, prerequisite dependencies, energy requirements, and impact with a 100% deterministic, explainable algorithm.
- "Your Next Best Task": Eliminates ambiguity by spotlighting the single highest-yield item to work on right now.
- Open-Weight Local AI: Ingests messy, natural-language…
Repository link: https://github.com/song-code-won/PrioritiAI
How I Built It
PrioritiAI was built around a hybrid architecture pairing local open-weight AI with deterministic execution:
1. Open-Source AI Engine (Ollama + Qwen 2.5 / Mistral)
-
Local Inference Runtime: Integrates directly with Ollama running locally at
http://localhost:11434. -
Open-Weight Models: Designed for
Qwen 2.5 (3B/7B)(Apache 2.0 license) andMistral 7B(Apache 2.0 license). - Task Parsing & Subtask Decomposition: The local model interprets conversational task dumps, identifies entities (durations, deadlines, energy levels), and decomposes complex tasks into 3–5 actionable subtasks with time estimates.
- Resilient Offline Fallback: If Ollama is unavailable or offline, PrioritiAI automatically engages an embedded heuristic rule engine. Zero features or priority calculations ever break.
2. Deterministic Prioritization Scoring Formula
Instead of relying on unpredictable black-box rankings, PrioritiAI uses a mathematically transparent priority formula ($0$ to $100$ pts):
$$\text{Priority Score} = \min\left(100, \max\left(0, \text{Base Score} - \text{Dependency Penalty}\right)\right)$$
Where:
$$\text{Base Score} = (\text{Urgency} \times 0.40) + (\text{Importance} \times 0.35) + (\text{Effort} \times 0.15) + (\text{Impact} \times 0.10) + \text{Energy Bonus}$$
-
Dependency Penalty (-80 pts): Tasks blocked by incomplete prerequisite tasks are penalized and moved to
Waiting on Something. - Quick Win Bonus (+40 pts): Sub-15 minute tasks receive an incentive boost to create early momentum.
3. Modern Tech Stack
- Frontend: React 19, TypeScript 5.9, Tailwind CSS.
- State & Storage: React Context + client-side LocalStorage persistence with full JSON export/import.
- Testing: Automated unit tests with Vitest covering priority scoring, blocker penalties, and the daily planner.
Why Does Open Innovation Matter?
Open innovation is what made PrioritiAI possible, practical, and ethically aligned with user trust:
Complete Data Sovereignty & Confidentiality:
With proprietary cloud APIs (OpenAI, Claude, etc.), every task—whether it's confidential client milestone notes or personal deadlines—must travel across the public internet to third-party data centers. Open-weight models running on consumer hardware ensure that zero bits of private task data ever leave the user's laptop.Zero Recurring API Bills for Side Hustlers:
For students and independent creators, paying $\$20$/month or per-token API charges for daily task management destroys the economics of a personal productivity tool. Open-weight models like Qwen 2.5 run on standard modern hardware at $\$0.00$ operational inference cost forever.Guaranteed Offline Resilience:
Productivity software cannot be brittle. Freelancers working on trains, flights, or with intermittent connections cannot afford to lose their task prioritizer. Open-source local AI delivers uninterrupted productivity anywhere on earth.Transparent, Explainable Logic:
Proprietary AI wrappers create black boxes where users don't know why a task was ranked where it was. By open-sourcing the deterministic scoring algorithm alongside local models, users have complete visibility and trust into why their next task was recommended.
Prize Categories
- Primary Challenge: Hacktoberfest Open-Source AI Challenge Week 2 — Side Hustle
- Open-Source AI & Local Inference Track
Top comments (0)