DEV Community: Shikhar Ahirwar

🚀 Soft Start — A Local AI Tool to Turn Overwhelming Tasks into Tiny First Steps

Shikhar Ahirwar — Tue, 20 Oct 2026 05:52:14 +0000

💡 What I Built

I built Soft Start 🌱 — a privacy-first local AI tool designed to help people start tasks they have been putting off.

Instead of generating a huge and overwhelming plan, Soft Start turns a task into one tiny first step 👣 and up to three optional next steps.

It also adapts the plan to your current energy level ⚡:

🔋 Low energy
⚡ Some energy
🚀 Plenty of energy

The AI runs locally using Ollama + an open-weight model 🤖, so your task stays on your device instead of being sent to a cloud AI service.

The goal is simple:

🌱 Make starting easier — not make the task more complicated.

⚙️ How It Works

📝 Enter a task you have been avoiding.
⚡ Select your current energy level.
🤖 Soft Start sends the task to your local Ollama model.
💡 The model creates one small first step and optional next steps.
👀 Review the suggested plan.
🚶 Start with the first step.

🔒 The AI only provides suggestions. It cannot execute the task or access your filesystem.

🎯 Why I Built It

Sometimes the hardest part of a task is simply getting started.

I wanted to experiment with open-source AI 🤖 in a way that was actually useful without making AI responsible for completing everything.

Soft Start focuses on reducing the starting barrier 🧩 while keeping the final decision and action with the user.

I also wanted to explore how local AI 🖥️ can provide a more private alternative for simple personal workflows.

🌱 One small step can be enough to get moving.

🔒 Privacy

Privacy was an important part of the design.

🏠 AI inference runs locally through Ollama.

🚫 No cloud AI API is required.

📊 No analytics.

🍪 No cookies.

💾 No database for task storage.

🌐 The Node.js server binds to 127.0.0.1.

💻 Code
🚀 https://github.com/shikharahirwar489-ui/soft-start-local-ai.git
GitHub

Your task stays on your device. 🔐

Quote Forge

Shikhar Ahirwar — Wed, 14 Oct 2026 16:32:02 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle

What I Built

I built QuoteForge Local, a private, local-first AI proposal and quote builder designed for freelancers.

A freelancer can enter:

A client's project brief
Service or project type
Hourly rate
Estimated hours
Additional expenses

QuoteForge then generates a professional proposal draft using the open-weight Llama 3.2 model running locally through Ollama.

The important part is that AI does not control the price.

The quote is calculated separately by application code:

Hourly Rate × Estimated Hours + Additional Expenses

This keeps the financial calculation deterministic while AI is used only for proposal wording.

Who would pay for it?

The target users are freelancers, independent developers, designers, marketers, consultants, and other solo service providers who regularly create client proposals.

The value proposition is simple: spend less time writing repetitive proposals while keeping sensitive client information local.

What problem does it solve?

Freelancers often have to repeatedly turn a short client brief into a structured proposal. Existing AI tools can help with the writing, but they often require sending client information to a hosted AI service.

QuoteForge explores a different approach:

Can AI help create useful freelance proposals without requiring the client's project brief to leave the freelancer's machine?

Demo

Repository:

https://github.com/shikharahirwar489-ui/quoteforge-local

The application currently runs locally with Docker Compose.

docker compose up --build

Then open:

http://127.0.0.1:8000

The first launch downloads the Ollama runtime/model. Later launches reuse the local model volume.

There is currently no hosted demo because the project is intentionally designed around local inference.

Code

GitHub Repository:

https://github.com/shikharahirwar489-ui/quoteforge-local

The repository contains the complete application, Docker configuration, frontend, backend, prompts, and setup instructions.

The project is released under the MIT License.

How I Built It

QuoteForge is built around open-weight AI running locally.

AI

Llama 3.2 — open-weight language model
Ollama — local model runtime

The model is responsible for generating the proposal wording from the freelancer's project information.

I deliberately kept the AI's responsibility narrow.

The model is instructed to avoid inventing:

Credentials
Previous clients
Deliverables
Deadlines
Guarantees
Other unsupported claims

If important information is missing, the proposal should flag it for the freelancer to confirm instead of making it up.

Application

The application

Quote Forge

Shikhar Ahirwar — Wed, 14 Oct 2026 16:26:31 +0000

QuoteForge Local

A private, local-first AI proposal and quote builder for freelancers.

QuoteForge Local helps freelancers turn a client project brief into a professional proposal draft while keeping the client's information and AI processing on their own machine.

Describe the project, enter your hourly rate, estimated hours, and additional expenses. QuoteForge then calculates the project estimate using ordinary application code and uses the open-weight Llama 3.2 model through Ollama to draft the proposal.

Built for the Hacktoberfest Open-Source AI Challenge — Week 2: Side Hustle.

✨ What It Does

QuoteForge Local combines deterministic pricing with local AI assistance:

📝 AI Proposal Drafting — Generate a proposal from a project brief.
💰 Deterministic Quote Calculation — Calculates:

Hourly Rate × Estimated Hours + Additional Expenses

🔒 Local-First AI — The project brief is processed through a local Ollama service instead of a hosted AI API.
🤖 Open-Weight AI — Uses Llama 3.2 through Ollama.
🚫 No API Keys Required — No OpenAI, Anthropic, or other hosted AI account is required.
📋 Copy-Ready Output — Review and copy the generated proposal directly from the browser.
🧠 Controlled AI Scope — The model is instructed not to invent credentials, deliverables, deadlines, guarantees, or other unsupported claims.
💾 No Database — Generated content stays in the browser until the user chooses to copy it.
📊 No Analytics — No tracking or third-party analytics are included.
📤 No Automatic Sending — QuoteForge does not send proposals directly to clients.

🔐 Why Local-First?

Client briefs can contain sensitive business information.

Instead of sending that information to a third-party AI API, QuoteForge keeps the AI workflow local:

Browser → Local Python API → Local Ollama → Llama 3.2

The application is designed so that the project brief is sent only to the locally running Ollama service.

The first Docker Compose startup requires internet access to download the Docker runtime components and the selected Ollama model. After the model has been downloaded, the model is reused from the local Docker volume.

🧮 AI Does Not Control the Price

One important design decision is separating language generation from financial calculation.

The AI is responsible for wording the proposal.

The application code is responsible for calculating the quote.

For example:

₹1,000/hour × 20 hours + ₹2,000 expenses = ₹22,000

The model cannot modify this arithmetic because the calculation happens independently in ordinary application code.

This makes the system more predictable and prevents an AI-generated response from accidentally changing the quoted amount.

🤖 Responsible AI Design

QuoteForge deliberately gives the language model a narrow responsibility.

The model generates proposal wording, but it does not:

Decide whether a project is profitable
Determine the freelancer's pricing
Create binding contract terms
Invent previous client experience
Invent qualifications or credentials
Promise unsupported results
Automatically send proposals to clients

Missing information should be identified for the freelancer to confirm rather than being fabricated.

AI-generated content should still be reviewed by the freelancer before being shared with a client.

🛠️ Tech Stack

Python 3.12 — Standard library
HTML / CSS / JavaScript — Frontend
Ollama — Local AI runtime
Llama 3.2 — Open-weight language model
Docker Compose — Local deployment
MIT License

The Python backend intentionally uses the standard library, including modules such as:

http.server
urllib

This keeps the project lightweight and reduces unnecessary dependencies.

🚀 Run Locally

Requirements

You need:

Docker Desktop
Docker Compose
Sufficient disk space for the Ollama image and model

Start QuoteForge

docker compose up --build

The first startup downloads the required Ollama model.

Once the services are ready, open:

http://127.0.0.1:8000

Stop the application with:

Ctrl+C

Later launches can reuse the locally stored model through the Docker volume.

⚙️ Changing the AI Model

The default setup uses:

llama3.2

The OLLAMA_MODEL configuration can be changed to another model available through Ollama, provided the machine has enough resources to run it.

🏗️ Architecture

                ┌──────────────────┐
                │     Browser      │
                │  HTML/CSS/JS     │
                └────────┬─────────┘
                         │
                         ▼
                ┌──────────────────┐
                │   Python API     │
                │  Local Adapter   │
                └───────┬──────────┘
                        │
              ┌─────────┴─────────┐
              │                   │
              ▼                   ▼
      ┌──────────────┐    ┌──────────────┐
      │ Quote Logic  │    │   Ollama     │
      │ Rate × Hours │    │  Llama 3.2   │
      │ + Expenses   │    │ Local Model  │
      └──────────────┘    └──────────────┘

The quote calculation and AI generation remain separate by design.

🎯 The Problem

Freelancers often need to create proposals quickly after receiving a client inquiry.

That process can involve:

Understanding the client's brief
Structuring the proposed work
Writing professional proposal language
Calculating the estimated price
Reviewing the proposal for unsupported claims
Sending it to the client

Many AI proposal tools require sending client information to a hosted AI service.

QuoteForge explores a different approach:

Can AI help freelancers write proposals without requiring their client data to leave their machine?

💡 The Solution

QuoteForge Local combines a simple deterministic quote calculator with a locally running open-weight language model.

The result is a small, transparent workflow:

Client Brief → Local AI Draft → Deterministic Quote → Human Review → Copy

The freelancer remains in control of the final proposal.

🌱 Why This Fits "Side Hustle"

QuoteForge is designed around a practical freelancer workflow rather than AI for AI's sake.

A freelancer can use it to quickly turn an initial project description into a structured proposal while maintaining control over:

Pricing
Scope
Client information
Final wording
When and where the proposal is sent

The goal is not to replace the freelancer.

The goal is to remove repetitive proposal-writing work while keeping important decisions under human control.

⚠️ Limitations

QuoteForge is intentionally small and focused.

It currently does not:

Store client accounts
Provide a CRM
Send emails
Sign contracts
Process payments
Determine whether a project is financially profitable
Guarantee the accuracy of AI-generated text
Provide legal or financial advice

The generated proposal should always be reviewed before being shared with a client.

🔮 Future Ideas

Possible future improvements include:

Multiple proposal templates
Export to PDF
Proposal version history stored locally
More detailed project breakdowns
Local proposal history
Additional Ollama model presets
Custom freelancer profiles stored locally
Optional local document import
Improved accessibility
Automated proposal quality checks

These features would be considered carefully to preserve the project's local-first philosophy.

🌍 Privacy Philosophy

QuoteForge follows a simple principle:

Your client data should not need to become someone else's AI training input just to help you write a proposal.

There are no third-party fonts, analytics services, hosted AI APIs, or accounts required by the application itself.

The initial Docker setup does require network access to obtain the container image and Ollama model. Once running locally, proposal generation uses the local Ollama service.

📜 License

QuoteForge Local is released under the MIT License.

See LICENSE for details.

🏆 Hacktoberfest

Built as a submission for:

Hacktoberfest Open-Source AI Challenge — Week 2: Side Hustle

QuoteForge demonstrates how open-weight AI can be integrated into a realistic freelance workflow while keeping pricing deterministic and AI processing local.

Repository: shikharahirwar489-ui/quoteforge-local

Project: QuoteForge Local

Focus: Local AI • Freelancing • Privacy • Open-Weight Models • Responsible AI

Quote Forge

Shikhar Ahirwar — Wed, 14 Oct 2026 16:16:54 +0000

QuoteForge Local

A private, local-first AI proposal and quote builder for freelancers.

QuoteForge Local helps freelancers turn a client project brief into a professional proposal draft while keeping the client's information and AI processing on their own machine.

Built for the Hacktoberfest Open-Source AI Challenge — Week 2: Side Hustle.

✨ What It Does

QuoteForge Local combines deterministic pricing with local AI assistance:

📝 AI Proposal Drafting — Generate a proposal from a project brief.
💰 Deterministic Quote Calculation — Calculates:

Hourly Rate × Estimated Hours + Additional Expenses

🔒 Local-First AI — The project brief is processed through a local Ollama service instead of a hosted AI API.
🤖 Open-Weight AI — Uses Llama 3.2 through Ollama.
🚫 No API Keys Required — No OpenAI, Anthropic, or other hosted AI account is required.
📋 Copy-Ready Output — Review and copy the generated proposal directly from the browser.
🧠 Controlled AI Scope — The model is instructed not to invent credentials, deliverables, deadlines, guarantees, or other unsupported claims.
💾 No Database — Generated content stays in the browser until the user chooses to copy it.
📊 No Analytics — No tracking or third-party analytics are included.
📤 No Automatic Sending — QuoteForge does not send proposals directly to clients.

🔐 Why Local-First?

Client briefs can contain sensitive business information.

Instead of sending that information to a third-party AI API, QuoteForge keeps the AI workflow local:

Browser → Local Python API → Local Ollama → Llama 3.2

The application is designed so that the project brief is sent only to the locally running Ollama service.

🧮 AI Does Not Control the Price

One important design decision is separating language generation from financial calculation.

The AI is responsible for wording the proposal.

The application code is responsible for calculating the quote.

For example:

₹1,000/hour × 20 hours + ₹2,000 expenses = ₹22,000

The model cannot modify this arithmetic because the calculation happens independently in ordinary application code.

This makes the system more predictable and prevents an AI-generated response from accidentally changing the quoted amount.

🤖 Responsible AI Design

QuoteForge deliberately gives the language model a narrow responsibility.

The model generates proposal wording, but it does not:

Decide whether a project is profitable
Determine the freelancer's pricing
Create binding contract terms
Invent previous client experience
Invent qualifications or credentials
Promise unsupported results
Automatically send proposals to clients

Missing information should be identified for the freelancer to confirm rather than being fabricated.

AI-generated content should still be reviewed by the freelancer before being shared with a client.

🛠️ Tech Stack

Python 3.12 — Standard library
HTML / CSS / JavaScript — Frontend
Ollama — Local AI runtime
Llama 3.2 — Open-weight language model
Docker Compose — Local deployment
MIT License

The Python backend intentionally uses the standard library, including modules such as:

http.server
urllib

This keeps the project lightweight and reduces unnecessary dependencies.

🚀 Run Locally

Requirements

You need:

Docker Desktop
Docker Compose
Sufficient disk space for the Ollama image and model

Start QuoteForge

docker compose up --build

The first startup downloads the required Ollama model.

Once the services are ready, open:

http://127.0.0.1:8000

Stop the application with:

Ctrl+C

Later launches can reuse the locally stored model through the Docker volume.

⚙️ Changing the AI Model

The default setup uses:

llama3.2

The OLLAMA_MODEL configuration can be changed to another model available through Ollama, provided the machine has enough resources to run it.

🏗️ Architecture

                ┌──────────────────┐
                │     Browser      │
                │  HTML/CSS/JS     │
                └────────┬─────────┘
                         │
                         ▼
                ┌──────────────────┐
                │   Python API     │
                │  Local Adapter   │
                └───────┬──────────┘
                        │
              ┌─────────┴─────────┐
              │                   │
              ▼                   ▼
      ┌──────────────┐    ┌──────────────┐
      │ Quote Logic  │    │   Ollama     │
      │ Rate × Hours │    │  Llama 3.2   │
      │ + Expenses   │    │ Local Model  │
      └──────────────┘    └──────────────┘

The quote calculation and AI generation remain separate by design.

🎯 The Problem

Freelancers often need to create proposals quickly after receiving a client inquiry.

That process can involve:

Understanding the client's brief
Structuring the proposed work
Writing professional proposal language
Calculating the estimated price
Reviewing the proposal for unsupported claims
Sending it to the client

Many AI proposal tools require sending client information to a hosted AI service.

QuoteForge explores a different approach:

Can AI help freelancers write proposals without requiring their client data to leave their machine?

💡 The Solution

QuoteForge Local combines a simple deterministic quote calculator with a locally running open-weight language model.

The result is a small, transparent workflow:

Client Brief → Local AI Draft → Deterministic Quote → Human Review → Copy

The freelancer remains in control of the final proposal.

🌱 Why This Fits "Side Hustle"

QuoteForge is designed around a practical freelancer workflow rather than AI for AI's sake.

A freelancer can use it to quickly turn an initial project description into a structured proposal while maintaining control over:

Pricing
Scope
Client information
Final wording
When and where the proposal is sent

The goal is not to replace the freelancer.

The goal is to remove repetitive proposal-writing work while keeping important decisions under human control.

⚠️ Limitations

QuoteForge is intentionally small and focused.

It currently does not:

Store client accounts
Provide a CRM
Send emails
Sign contracts
Process payments
Determine whether a project is financially profitable
Guarantee the accuracy of AI-generated text
Provide legal or financial advice

The generated proposal should always be reviewed before being shared with a client.

🔮 Future Ideas

Possible future improvements include:

Multiple proposal templates
Export to PDF
Proposal version history stored locally
More detailed project breakdowns
Local proposal history
Additional Ollama model presets
Custom freelancer profiles stored locally
Optional local document import
Improved accessibility
Automated proposal quality checks

These features would be considered carefully to preserve the project's local-first philosophy.

🌍 Privacy Philosophy

QuoteForge follows a simple principle:

Your client data should not need to become someone else's AI training input just to help you write a proposal.

There are no third-party fonts, analytics services, hosted AI APIs, or accounts required by the application itself.

The initial Docker setup does require network access to obtain the container image and Ollama model. Once running locally, proposal generation uses the local Ollama service.

📜 License

QuoteForge Local is released under the MIT License.

See LICENSE for details.

🏆 Hacktoberfest

Built as a submission for:

Hacktoberfest Open-Source AI Challenge — Week 2: Side Hustle

QuoteForge demonstrates how open-weight AI can be integrated into a realistic freelance workflow while keeping pricing deterministic and AI processing local.

Repository: shikharahirwar489-ui/quoteforge-local

Project: QuoteForge Local

Focus: Local AI • Freelancing • Privacy • Open-Weight Models • Responsible AI

QuoteForge: Local AI proposals with prices you can trust

Shikhar Ahirwar — Wed, 14 Oct 2026 15:51:19 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle.

What I built

QuoteForge helps an independent freelancer turn a client brief into a clear proposal. Enter the service, brief, hourly rate, estimated hours, and expenses; the app calculates the quote and asks a local open-weight model to draft the words around it.

Demo

The demo runs locally. Clone the repository, then run:

docker compose up --build

After the first-run download of the Ollama runtime and llama3.2 model completes, open http://127.0.0.1:8000. The first download needs disk space and can take a while. There is no hosted demo yet.

Code

QuoteForge on GitHub

How I built it

The app uses a small Python standard-library HTTP server, a plain HTML/CSS/JavaScript interface, Docker Compose, and Ollama running the open-weight llama3.2 model. The model drafts the proposal; ordinary code calculates rate × hours + expenses, so generated text cannot alter the estimate.

The prompt asks the model not to invent experience, deliverables, dates, or guarantees. Missing details are marked for the freelancer to confirm. The draft stays in the browser until copied; the app has no account, database, analytics, or send-to-client feature.

Why open innovation matters

A client brief can include private business details. Running the model locally keeps that brief on the freelancer’s machine and avoids sending it to a hosted AI API. The model can also be changed through the OLLAMA_MODEL setting. The trade-off is that the user supplies local compute and manages the initial model download.

I kept the first version deliberately narrow: it drafts wording, while the freelancer reviews the scope and decides whether the price is right. That matches practical guidance from Shawn M. on reliability and narrow AI workflows and bestbee on gathering verifiable evidence when evaluating open-source AI tools.

Limitations

QuoteForge has not been tested with a real client workflow yet. Model output can still be wrong, so review it before sharing. The current demo is local-only and requires Docker plus enough disk space for the model.

GO. — AI for Leaving the Screen

Shikhar Ahirwar — Thu, 08 Oct 2026 14:26:23 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

I Built an AI That Wants You to Put Your Phone Away 🌱

What I Built

GO. — AI for Leaving the Screen is a small web app built around a simple idea:

Ask AI. Then leave.

Most AI tools are designed to keep us on screens longer. GO. tries to do the opposite.

The user chooses a vibe, duration, company, and place, and GO. generates a small offline mission designed to get them away from their device and into the real world.

Examples include:

Take a short walk and notice five things you normally ignore.
Explore a nearby area you've never paid attention to.
Do a small creative activity without using your phone.
Talk to someone you haven't spoken to recently.
Spend 20 minutes outside without checking your screen.

After generating a mission, GO. starts a short countdown and tells the user to put the phone away.

The app is designed for students, people who spend too much time online, and anyone who wants a simple reason to step outside, move, notice their surroundings, or connect with other people.

The goal isn't to replace real life with AI.

The goal is to use AI for a few seconds so you can leave it behind.

Demo

🌱 Live Demo:
https://shikharahirwar489-ui.github.io/-GO.-AI-for-Leaving-the-Screen/

Code

💻 GitHub Repository:
https://github.com/shikharahirwar489-ui/-GO.-AI-for-Leaving-the-Screen.git

The project is built with a simple frontend so that it is easy for other developers to understand, modify, and contribute to.

How I Built It

GO. is built with HTML, CSS, and JavaScript on the frontend.

The project is designed around local/open AI rather than depending on a proprietary AI API.

The planned architecture is:

Browser → Local Backend → Ollama → Open-Weight Model → Offline Mission

The AI receives structured information such as:

selected activity/vibe
available time
whether the user is alone or with others
preferred location
optional constraints

It then generates a short, practical mission that encourages the user to stop using their device.

The frontend also includes:

responsive UI
mission generation
multiple activity categories
countdown before leaving the screen
mission completion
local mission history
localStorage
safety guidance
no account requirement

The important design decision is that AI is only the starting point. The experience intentionally ends by encouraging the user to stop interacting with the application.

Why Does Open Innovation Matter?

Open innovation makes an idea like GO. much more interesting because the AI doesn't have to live inside a closed ecosystem.

With open-weight models and local inference, the mission-generation system can potentially run directly on the user's own computer.

That means developers can:

inspect how the AI is being used
experiment with different models
modify the prompts
run inference locally
build new mission-generation strategies
contribute improvements without depending entirely on a proprietary API

It also fits the philosophy of the project.

The AI should be a tool, not another platform trying to capture the user's attention.

Open AI ecosystems make it possible to experiment with that idea while keeping the project transparent and hackable.

My Agent Session

I used AI-assisted development while building GO. to help design the application structure, frontend logic, UI, mission-generation system, and project documentation.

I have not included a DevRelay agent session link because I don't want to claim a session that isn't actually available.

Prize Categories

Hacktoberfest Open-Source AI Challenge — Week 1: Touch Grass

This project is primarily entering the Touch Grass challenge category.

The project focuses directly on the challenge's idea of using open-source/open-weight AI to encourage people to spend less time on screens and more time in the physical world.

Final Thought

We keep asking:

"What can AI help us do online?"

GO. asks a different question:

"What can AI help us do offline?"

Ask AI.

Get your mission.

Put the phone away.

🌱

🌱 GO. — An AI That Wants You to Stop Using It | Touch Grass Challenge

Shikhar Ahirwar — Wed, 07 Oct 2026 05:41:46 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

I Built an AI That Wants You to Put Your Phone Away 🌱

What I Built

GO. — AI for Leaving the Screen is a small web app built around a simple idea:

Ask AI. Then leave.

Most AI tools are designed to keep us on screens longer. GO. tries to do the opposite.

The user chooses a vibe, duration, company, and place, and GO. generates a small offline mission designed to get them away from their device and into the real world.

Examples include:

Take a short walk and notice five things you normally ignore.
Explore a nearby area you've never paid attention to.
Do a small creative activity without using your phone.
Talk to someone you haven't spoken to recently.
Spend 20 minutes outside without checking your screen.

After generating a mission, GO. starts a short countdown and tells the user to put the phone away.

The app is designed for students, people who spend too much time online, and anyone who wants a simple reason to step outside, move, notice their surroundings, or connect with other people.

The goal isn't to replace real life with AI.

The goal is to use AI for a few seconds so you can leave it behind.

Demo

🌱 Live Demo:
https://shikharahirwar489-ui.github.io/-GO.-AI-for-Leaving-the-Screen/

Code

💻 GitHub Repository:
https://github.com/shikharahirwar489-ui/-GO.-AI-for-Leaving-the-Screen.git

The project is built with a simple frontend so that it is easy for other developers to understand, modify, and contribute to.

How I Built It

GO. is built with HTML, CSS, and JavaScript on the frontend.

The project is designed around local/open AI rather than depending on a proprietary AI API.

The planned architecture is:

Browser → Local Backend → Ollama → Open-Weight Model → Offline Mission

The AI receives structured information such as:

selected activity/vibe
available time
whether the user is alone or with others
preferred location
optional constraints

It then generates a short, practical mission that encourages the user to stop using their device.

The frontend also includes:

responsive UI
mission generation
multiple activity categories
countdown before leaving the screen
mission completion
local mission history
localStorage
safety guidance
no account requirement

The important design decision is that AI is only the starting point. The experience intentionally ends by encouraging the user to stop interacting with the application.

Why Does Open Innovation Matter?

Open innovation makes an idea like GO. much more interesting because the AI doesn't have to live inside a closed ecosystem.

With open-weight models and local inference, the mission-generation system can potentially run directly on the user's own computer.

That means developers can:

inspect how the AI is being used
experiment with different models
modify the prompts
run inference locally
build new mission-generation strategies
contribute improvements without depending entirely on a proprietary API

It also fits the philosophy of the project.

The AI should be a tool, not another platform trying to capture the user's attention.

Open AI ecosystems make it possible to experiment with that idea while keeping the project transparent and hackable.

My Agent Session

I used AI-assisted development while building GO. to help design the application structure, frontend logic, UI, mission-generation system, and project documentation.

I have not included a DevRelay agent session link because I don't want to claim a session that isn't actually available.

Prize Categories

Hacktoberfest Open-Source AI Challenge — Week 1: Touch Grass

This project is primarily entering the Touch Grass challenge category.

The project focuses directly on the challenge's idea of using open-source/open-weight AI to encourage people to spend less time on screens and more time in the physical world.

Final Thought

We keep asking:

"What can AI help us do online?"

GO. asks a different question:

"What can AI help us do offline?"

Ask AI.

Get your mission.

Put the phone away.

🌱