This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine
💻 Code
GitHub Repository: https://github.com/amritayadav6738bosscoder-boop/dayscan
The repository is the starting point for DayScan. I will add the application code, setup instructions, and AI integration details as the project develops.
🛠️ How I Built It
My starting point was a simple idea: make daily task tracking easier and more engaging.
The proposed implementation uses:
DayScan: The daily-task tracking application.
Gemma: An open-weight language model for generating task summaries and helpful suggestions.
Ollama: A way to run a supported model locally.
Local AI inference: To help keep personal task information on the user's device, depending on the final implementation.
The intended workflow is straightforward:
The user adds tasks for the day.
DayScan stores and displays the task list.
The AI assistant reviews the task information provided to it.
The assistant generates a concise summary or suggests a next step.
The user decides what to do next.
I want to keep the experience simple and ensure that the AI supports the user's decisions rather than taking control of their schedule.
🌍 Why Does Open Innovation Matter?
Daily tasks can reveal a lot about someone's personal life, studies, work, and responsibilities.
Open-source AI offers an opportunity to build tools that are more transparent and adaptable. With a suitable locally running model, DayScan could help users process their task information without sending it to an external AI service.
Open models also give developers the flexibility to experiment with different models, study how the system behaves, and adapt the implementation to their needs.
For me, open innovation means being able to learn, experiment, and build useful technology without depending entirely on a closed AI platform.
I also want DayScan to be a project that other developers can understand, improve, and extend.
🎃 My Interpretation of “Ghost in the Machine”
The ghost represents the tasks we forget, the work we postpone, and the little reminders that follow us through the day.
Instead of making that ghost frightening, I want to make it helpful.
DayScan gives unfinished tasks a voice—not to make users feel guilty, but to help them regain focus and finish their day with greater clarity.
The spooky theme is fun, but the underlying goal is practical: help people remember what matters and take the next small step.
🚀 What's Next?
My next steps are to:
Build and test the core daily-task interface.
Integrate a working open-weight AI model.
Test local inference and document its setup requirements.
Add sample tasks and a demonstration workflow.
Document the architecture and instructions in the GitHub repository.
I’m excited to keep learning, experimenting, and turning this idea into a useful open-source project.
🏷️ Tags
devchallenge #hf26challenge #opensource #ai #gemma
hacktoberfest # devchallange
**_[](
Top comments (1)
Damm good