FixIt Forward — From Noticing a Problem to Taking Action
The problem isn't that we don't notice what's broken. It's that we don't always know what to do about it.
An overflowing dustbin. A broken streetlight. A damaged pavement. A neglected public space.
We encounter these problems in our everyday lives, but reporting them can be confusing. We might not know which authority to contact, how to describe the issue, or what information to include.
So, for the Hacktoberfest Open-Source AI Challenge — Week 1: Touch Grass, I'm building FixIt Forward — a project designed to turn everyday observations into structured, actionable civic reports.
The idea is simple: instead of noticing a public problem and walking past it, make it easier to take the next step toward getting it addressed.
What I Built
FixIt Forward is centered around one focused workflow: photograph a public issue, turn it into a clear report, and make it easier to act on it.
The intended workflow includes:
- Photo-based issue identification: Use AI to identify issues such as overflowing bins, damaged roads, broken streetlights, and neglected public spaces.
- AI-assisted report generation: Convert the observation into a concise, structured description.
- Location support: Allow users to add the location of the issue.
- Editable reports: Let users review and correct AI-generated information before using it.
- Report export: Make the report easy to copy or export for submission to the relevant civic authority.
- A privacy-conscious approach: Explore self-hosted, open-weight AI models so that image processing can happen locally when the chosen model and deployment support it.
The goal isn't to replace civic authorities or promise that every reported issue will be resolved.
It's to reduce the friction between noticing a problem and doing something about it.
Less scrolling past problems. More taking action in the real world.
Demo
I haven't deployed a live demo yet.
🚧 Project Status: Source repository available; deployment and a live walkthrough are not yet available.
Code
💻 GitHub Repository:
FixIt Forward — Source Code
How I Built It
I approached FixIt Forward as an application of open-source AI to a practical, everyday problem.
The project is designed around a straightforward pipeline:
Photo → Issue Classification → Report Generation → User Review → Export
The AI component is intended to help identify the type of public issue and draft a useful description. The user remains in control of the final report, including its wording and location.
I also want to explore open-weight models that can be self-hosted rather than making the application dependent on a proprietary AI API. This could offer communities greater control over their data and deployment, depending on the model and infrastructure used.
The focus is not simply on generating text with AI. It is on producing something useful outside the application: a report that a person can review and use to contact the appropriate authority.
Why Does Open Innovation Matter?
Public infrastructure affects everyone, but the tools available to report problems are not always straightforward or accessible.
Open innovation creates an opportunity to build tools that communities can adapt to their own needs.
A workflow developed for reporting overflowing bins in one city could be adapted to report damaged pavements, broken streetlights, or neglected public spaces elsewhere.
Using open-source AI could also make it possible to self-host parts of the workflow, inspect the implementation, and adapt the technology to local requirements instead of relying entirely on a closed service.
For me, the value of open innovation is making technology more accessible, adaptable, and useful beyond a single application.
The aim is not to build another AI chatbot. It's to use AI to help people take practical action in the physical world.
My Agent Session
I don't currently have a saved, shareable DevRelay agent session for FixIt Forward.
I'll leave this section without a session link rather than include an unrelated session or an unverified link.
Prize Categories
Hacktoberfest Open-Source AI Challenge — Week 1: Touch Grass
Final Thoughts
Touch Grass is a reminder to step away from our screens and engage with the world around us.
FixIt Forward takes that idea one step further: what if the things we notice while walking around our neighbourhoods became opportunities to make those places better?
A photograph won't fix a broken streetlight by itself. But a clear, actionable report can help turn an observation into a concrete next step.
Notice it. Report it. Fix it forward. 🌱
Top comments (0)