"To-Do Lists create noise. Awerlain creates focus."
Awerlain is not just another task manager. It is an Execution Intelligence System powered by Gemini 3, designed to bridge the gap between "Intention" and "Action." Built specifically for the Google DeepMind Gemini 3 Hackathon.
Modern knowledge workers face a crisis not of tools, but of cognitive bandwidth:
- Cognitive Overload: We have too many tasks and decision fatigue sets in before we even start.
- The Action Gap: We write tasks down ("Write Report"), but they feel too heavy to start, leading to procrastination.
- Friction: Manually inputting details, deadlines, and subtasks is tedious.
Result: To-do lists become graveyards of good intentions.
Awerlain uses Behavioral Psychology principles combined with Multimodal Generative AI to remove friction and force execution.
Stop typing out schedules or transcribing whiteboard notes.
- How it works: Upload a screenshot of a chat, a photo of a handwritten schedule, or an email thread.
- AI Power: Gemini Vision analyzes the image, extracts context/deadlines, and structures it into actionable data instantly.
Eliminate "Decision Paralysis." You don't choose what to do; the system does.
- How it works: Our engine calculates a dynamic Priority Score based on Deadline Proximity, Task Effort, User Energy Level, and Persona (e.g., Developer vs. Student).
- Result: The UI hides everything else and shows you only the ONE most important task to focus on right now.
A UI designed to kill distraction.
- How it works: When you start a task, the interface transforms into a full-screen Execution Mode. No sidebar, no navigation, no other tasks. Just a timer and the current micro-step.
Trying to skip a hard task? Not so fast.
- How it works: If you try to "Skip" a high-priority task, the AI Coach intercepts.
- AI Power: It asks why (e.g., "Too hard?", "Low energy?"). Gemini then acts as a Behavioral Scientist, generating a specific "Micro-Intervention" to persuade you to start (e.g., "Forget the quality, just write the first sentence for 2 minutes").
We leverage Gemini 2.0 Flash / Pro to power the 3 core layers of the system:
| System Layer | Function | Prompt Engineering Strategy |
|---|---|---|
| Input Layer | extractTasksFromImage |
Multimodal analysis to detect context types (Schedule vs. Chat vs. Notes) and extract ISO dates from pixels. |
| Logic Layer | breakdownTask |
Context-Aware Breakdown. It intelligently decides when to break a task down. If Effort Score > 3, it generates atomic micro-steps. If the task is simple, it saves tokens. |
| Behavior Layer | getAiIntervention |
Persona-based prompting. Gemini adopts the persona of a "Performance Coach" to generate psychological nudges instead of generic advice. |
Follow these steps to run Awerlain locally:
- Node.js 18+
- Database (SQLite for dev, PostgreSQL for prod)
- Google Gemini API Key
-
Clone the Repository
git clone [https://github.com/username/awerlain.git](https://github.com/username/awerlain.git) cd awerlain -
Install Dependencies
npm install
-
Setup Environment Create a
.envfile in the root directory:DATABASE_URL="file:./dev.db" GEMINI_API_KEY="paste_your_google_ai_studio_key_here"
-
Setup Database
npx prisma db push npx prisma generate
-
Run the Server
npm run dev
Open
http://localhost:3000in your browser.
- MVP: Smart Capture, Focus Card, Execution Mode.
- AI Integration: Vision Capture & Behavioral Coach.
- Phase 2: Voice Command Integration (Multimodal Audio).
- Phase 3: Calendar Bi-directional Sync.
- Phase 4: Team/Multi-player Mode.
This project was built for a hackathon, but we welcome contributions! Please feel free to open an Issue or Pull Request.
License: MIT Created by: Fatiya Labibah