This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady
What I Built
Ever have so many things on your mind that organizing them becomes a task of its own?
Ideas, reminders, unfinished work, and random thoughts pile up. You write them down, but turning that messy brain dump into a clear list of actionable tasks takes time and effort.
That's why I'm building wasdr — DumpToDo: a lightweight task-management application that helps turn unstructured thoughts into organized tasks with the help of local AI.
The idea is simple: dump your thoughts first, organize them second, and focus on getting things done.
What it does
Brain dumps: Capture thoughts without having to organize everything upfront.
AI-assisted task processing: Use a locally running language model to help structure unorganized input into tasks.
Task organization: Work with priorities, categories, time estimates, and subtasks.
Export: Export task information in Markdown and Logseq-compatible Markdown formats.
Local AI integration: Connect to Ollama and use a model installed on your computer, such as qwen3:4b.
Standalone HTML edition: Run the bundled application locally, subject to its runtime, configuration, and dependency requirements.
The goal is to reduce the mental overhead between thinking about something and knowing what to do next.
Instead of spending more time organizing your thoughts, you can spend more time acting on them.
Demo
1.workspace
2.markdown export
3.offline app
& more ones ! ...
Code
github repo
tarun-sane
/
Wasdr_dumptodo
wasdr is an offline-first, brain-dump-to-task management app designed to turn messy notes into actionable tasks. Instead of manually breaking down every thought, you can dump your thoughts into one place and let the app organize them into a structured task list.
| title | wasdr — DumpToDo: Turn Brain Dumps into Actionable Tasks with Local AI |
|---|---|
| tags | devchallenge, hf26challenge, ollama, opensource , gemma |
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady
What I Built
Ever had a head full of tasks, ideas, reminders, and things you need to get done — but organizing all of them feels like another task?
That's the problem I wanted to solve with wasdr — DumpToDo.
wasdr is an offline-first, brain-dump-to-task management app designed to turn messy notes into actionable tasks. Instead of manually breaking down every thought, you can dump your thoughts into one place and let the app organize them into a structured task list.
The goal is simple: spend less time managing your to-do list and more time actually doing the work.
What makes it interesting?
- Brain dumps to tasks: Turn unstructured notes into individual, actionable tasks.
- …
The project is packaged as a standalone HTML file, with the interface and application logic bundled into the file.
Project: wasdr — DumpToDo
Repository: A public repository .
How I Built It
I built wasdr around a simple workflow:
- Capture unstructured thoughts in a brain dump.
- Process the text into individual tasks.
- Organize those tasks with priorities, categories, estimates, and subtasks.
- Review the results and track completion.
- Export the organized work when needed.
- The application includes a local processing path and an integration with Ollama for local model inference.
The local AI setup
Ollama provides the local model runtime. The application connects to the Ollama API at:
My local setup includes the qwen3:4b model , and the gemma4:e4b-mlx. The application also supports selecting a model through its configuration.
run this command in terminal , to get use of Ollama in the app :-
'launchctl setenv OLLAMA_ORIGINS "" && killall Ollama 2>/dev/null; killall ollama 2>/dev/null; OLLAMA_ORIGINS="" ollama serve'
Because a browser page and a local Ollama server have different origins, browser access may require configuring Ollama's allowed origins using OLLAMA_ORIGINS.
Once the model is downloaded and the local service is configured, inference can run on the user's own machine without sending prompts to a hosted model API.
The standalone application also includes local task-processing logic, so its task organization workflow is not dependent on a remote AI service for every operation.
Why Does Open Innovation Matter?
Open innovation makes it easier to experiment with AI models, inspect implementations, and build workflows around the tools that work for you.
For this project, using Ollama and a locally runnable model makes the AI integration more flexible. I can experiment with model selection and local inference without tying the core workflow to one hosted inference provider.
It also creates an opportunity to learn about the practical details of local AI integration: server configuration, browser-origin restrictions, model availability, and fallback behavior.
Most importantly, it lets me explore a productivity use case with accessible AI tooling.
I don't want AI to make every decision for the user. I want it to make the transition from an unstructured thought to a useful next action easier.
The Lazy Payoff
The payoff I'm aiming for is simple: less time organizing thoughts manually and less friction when deciding what to do next.
The intended workflow is:
Before: A long, unstructured brain dump that needs manual sorting.
After: A set of task suggestions that can be reviewed, prioritized, and organized.
I don't have verified time-saving measurements or benchmark results yet, so I won't invent percentages or claim a specific number of hours saved.
For now, the project explores one practical idea: let me capture thoughts quickly and use local AI to help turn them into something actionable.
Prize Categories
Best use of Gemma: to run and used to power the task creation from messy of notes ! :)
Top comments (0)