DEV Community: adarsh pandey

localBiz-ai

adarsh pandey — Sun, 18 Oct 2026 11:28:14 +0000

💼 LocalBiz AI: A Private, Open-Source AI Business Assistant for Small Businesses

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

What I Built

What if small business owners could use AI to run their businesses without paying for another monthly AI subscription or sending sensitive business information to a cloud AI service?

That question inspired me to build LocalBiz AI, an open-source AI business assistant designed for independent sellers, freelancers, and solo entrepreneurs.

Small business owners often spend hours writing product descriptions, responding to customer inquiries, and creating promotional content. While AI can automate much of this work, subscription costs and data privacy concerns can make cloud-based solutions less attractive for small businesses.

LocalBiz AI takes a different approach: use open-weight AI running locally to make everyday business tasks more accessible, private, and affordable.

✨ Key Features

🛍️ Product Listing Generator: Create tailored product descriptions for Etsy, Shopify, eBay, and social media, with customizable tones and word lengths.
💬 Customer Support Assistant: Draft professional customer replies using business-provided information and policies, helping avoid unsupported promises and invented discounts.
📢 Promotional Content Generator: Generate marketing copy and social media captions based on verified promotional terms.
🏪 Business Profiles: Save brand voice, preferred language, and standard policies using SQLite.
📚 Generation History: Review previous outputs, copy content, restore generation parameters, and delete individual records.
🔒 Local-First AI: Run inference through Ollama on the user's own computer instead of relying on a paid cloud AI API.
🛡️ Prompt Injection Safeguards: Treat customer messages as untrusted input and separate them from the instructions governing business policies.
📊 System Diagnostics: Check local Ollama availability and loopback configuration through the dashboard.

🎯 Who Is It For?

LocalBiz AI is designed for independent online sellers, home-based businesses, freelancers, and entrepreneurs who need practical AI assistance without adding another mandatory monthly AI subscription.

The potential business model includes paid installation, customization, and support services while keeping the core application open-source.

My goal is to explore how local AI can help small businesses save time on repetitive writing tasks while maintaining control over their business information.

Demo

🎬 Video Demo: [https://youtu.be/tJZip9FdcdY]

The repository includes a five-minute demonstration script at docs/demo-script.md, covering the application's key features and technical implementation.

The application currently runs locally using FastAPI and Ollama.

Code

💻 GitHub Repository:
https://github.com/adarshjnpandey-sys/localbiz-ai

License: MIT License for the application code; Meta Llama 3.2 Community License for the model weights.

The repository includes the application source code, frontend dashboard, backend APIs, automated tests, setup instructions, architecture documentation, and security considerations.

How I Built It

LocalBiz AI is built around open-weight AI inference running locally, rather than a proprietary cloud AI API.

Technology Stack

AI Model: Meta Llama 3.2 3B (llama3.2:3b)
Inference Engine: Ollama
Backend: Python, FastAPI, Pydantic v2
HTTP Client: HTTPX for asynchronous communication with Ollama
Database: SQLite
Frontend: HTML5, CSS3, and vanilla JavaScript ES6
Testing: Automated tests, including integration tests using the real local model

How It Works

The user selects a business tool and provides the relevant information.
The FastAPI backend validates the request and prepares the model instructions.
The backend communicates with the locally running Ollama service.
Llama 3.2 3B generates the requested business content.
The frontend displays the result, while relevant business profiles and generation history are stored locally.

Engineering and Reliability

Beyond basic text generation, I implemented:

Configurable model selection.
Structured request validation.
Error handling for Ollama downtime, request timeouts, and empty responses.
Prompt constraints designed to reduce unsupported business claims.
Separation of untrusted customer input from business instructions.
Persistent business profiles and generation history.
Automated backend and live-model integration tests.

The project currently includes 32 automated tests, with all 32 reported passing.

Documented generation latency ranges from approximately 6.8 to 14.8 seconds under the tested conditions.

The frontend uses plain HTML, CSS, and JavaScript, avoiding a heavyweight build pipeline.

Why Does Open Innovation Matter?

Open innovation is the foundation of LocalBiz AI.

Small businesses should have the opportunity to experiment with AI without automatically committing to recurring cloud inference charges or sending every prompt to an external provider.

Using an open-weight model through Ollama makes several things possible:

💰 Affordable experimentation: Users can run inference on compatible hardware they already own, avoiding mandatory per-request cloud AI charges. Electricity, hardware, and maintenance still have costs.

🔐 Greater data control: Local inference allows business prompts and information to stay on the user's machine during normal operation, provided the application and environment are configured accordingly.

🧩 Flexibility: Developers can configure the model and extend the application without being locked into a single proprietary AI provider.

🌍 Accessibility: Independent entrepreneurs can experiment with AI-assisted business workflows without needing a large software budget.

🤝 Community innovation: The MIT-licensed application code can be inspected, modified, and extended by other developers, subject to applicable third-party licenses.

For this project, open innovation is not simply about making code available. It is about giving small businesses more control over their technology, costs, and data.

My First User

LocalBiz AI is built around the needs of small online sellers, freelancers, and solo entrepreneurs.

The repository includes docs/customer-validation.md, which outlines target customer personas, discovery interview questions, usability-testing protocols, and a methodology for measuring time saved.

The next step is to validate the application with real business owners, gather feedback, and determine which features provide enough value for customers to pay for.

Prize Categories

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

🚀 Final Thoughts

I built LocalBiz AI around a simple idea:

Small businesses deserve access to useful AI tools without automatically accepting another monthly subscription or giving up control over their business information.

By combining open-weight AI, local inference, practical business workflows, and an open-source architecture, LocalBiz AI explores a more accessible approach to AI-powered business assistance.

The application is built. Now, the next challenge is to put it in front of real business owners, learn from their feedback, and make it genuinely useful.

Built with Python, FastAPI, Ollama, and Llama 3.2 3B.

🌱 GitHub: https://github.com/adarshjnpandey-sys/localbiz-ai

Hacktoberfest #OpenSource #ArtificialIntelligence #Llama #Ollama #SideHustle #SmallBusiness #Python

localbiz-ai:ai assitant for small bussiness

adarsh pandey — Sun, 18 Oct 2026 11:25:16 +0000

💼 LocalBiz AI: A Private, Open-Source AI Business Assistant for Small Businesses

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

What I Built

What if small business owners could use AI to run their businesses without paying for another monthly AI subscription or sending sensitive business information to a cloud AI service?

That question inspired me to build LocalBiz AI, an open-source AI business assistant designed for independent sellers, freelancers, and solo entrepreneurs.

LocalBiz AI takes a different approach: use open-weight AI running locally to make everyday business tasks more accessible, private, and affordable.

✨ Key Features

🛍️ Product Listing Generator: Create tailored product descriptions for Etsy, Shopify, eBay, and social media, with customizable tones and word lengths.
💬 Customer Support Assistant: Draft professional customer replies using business-provided information and policies, helping avoid unsupported promises and invented discounts.
📢 Promotional Content Generator: Generate marketing copy and social media captions based on verified promotional terms.
🏪 Business Profiles: Save brand voice, preferred language, and standard policies using SQLite.
📚 Generation History: Review previous outputs, copy content, restore generation parameters, and delete individual records.
🔒 Local-First AI: Run inference through Ollama on the user's own computer instead of relying on a paid cloud AI API.
🛡️ Prompt Injection Safeguards: Treat customer messages as untrusted input and separate them from the instructions governing business policies.
📊 System Diagnostics: Check local Ollama availability and loopback configuration through the dashboard.

🎯 Who Is It For?

The potential business model includes paid installation, customization, and support services while keeping the core application open-source.

My goal is to explore how local AI can help small businesses save time on repetitive writing tasks while maintaining control over their business information.

Demo

🎬 Video Demo: [https://youtu.be/tJZip9FdcdY]

The repository includes a five-minute demonstration script at docs/demo-script.md, covering the application's key features and technical implementation.

The application currently runs locally using FastAPI and Ollama.

Code

💻 GitHub Repository:
https://github.com/adarshjnpandey-sys/localbiz-ai

License: MIT License for the application code; Meta Llama 3.2 Community License for the model weights.

The repository includes the application source code, frontend dashboard, backend APIs, automated tests, setup instructions, architecture documentation, and security considerations.

How I Built It

LocalBiz AI is built around open-weight AI inference running locally, rather than a proprietary cloud AI API.

Technology Stack

AI Model: Meta Llama 3.2 3B (llama3.2:3b)
Inference Engine: Ollama
Backend: Python, FastAPI, Pydantic v2
HTTP Client: HTTPX for asynchronous communication with Ollama
Database: SQLite
Frontend: HTML5, CSS3, and vanilla JavaScript ES6
Testing: Automated tests, including integration tests using the real local model

How It Works

The user selects a business tool and provides the relevant information.
The FastAPI backend validates the request and prepares the model instructions.
The backend communicates with the locally running Ollama service.
Llama 3.2 3B generates the requested business content.
The frontend displays the result, while relevant business profiles and generation history are stored locally.

Engineering and Reliability

Beyond basic text generation, I implemented:

Configurable model selection.
Structured request validation.
Error handling for Ollama downtime, request timeouts, and empty responses.
Prompt constraints designed to reduce unsupported business claims.
Separation of untrusted customer input from business instructions.
Persistent business profiles and generation history.
Automated backend and live-model integration tests.

The project currently includes 32 automated tests, with all 32 reported passing.

Documented generation latency ranges from approximately 6.8 to 14.8 seconds under the tested conditions.

The frontend uses plain HTML, CSS, and JavaScript, avoiding a heavyweight build pipeline.

Why Does Open Innovation Matter?

Open innovation is the foundation of LocalBiz AI.

Small businesses should have the opportunity to experiment with AI without automatically committing to recurring cloud inference charges or sending every prompt to an external provider.

Using an open-weight model through Ollama makes several things possible:

🧩 Flexibility: Developers can configure the model and extend the application without being locked into a single proprietary AI provider.

🌍 Accessibility: Independent entrepreneurs can experiment with AI-assisted business workflows without needing a large software budget.

🤝 Community innovation: The MIT-licensed application code can be inspected, modified, and extended by other developers, subject to applicable third-party licenses.

For this project, open innovation is not simply about making code available. It is about giving small businesses more control over their technology, costs, and data.

My First User

LocalBiz AI is built around the needs of small online sellers, freelancers, and solo entrepreneurs.

The repository includes docs/customer-validation.md, which outlines target customer personas, discovery interview questions, usability-testing protocols, and a methodology for measuring time saved.

The next step is to validate the application with real business owners, gather feedback, and determine which features provide enough value for customers to pay for.

Prize Categories

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

🚀 Final Thoughts

I built LocalBiz AI around a simple idea:

Small businesses deserve access to useful AI tools without automatically accepting another monthly subscription or giving up control over their business information.

By combining open-weight AI, local inference, practical business workflows, and an open-source architecture, LocalBiz AI explores a more accessible approach to AI-powered business assistance.

The application is built. Now, the next challenge is to put it in front of real business owners, learn from their feedback, and make it genuinely useful.

Built with Python, FastAPI, Ollama, and Llama 3.2 3B.

🌱 GitHub: https://github.com/adarshjnpandey-sys/localbiz-ai

Hacktoberfest #OpenSource #ArtificialIntelligence #Llama #Ollama #SideHustle #SmallBusiness #Python

Green_ai:Screen time reduction

adarsh pandey — Fri, 09 Oct 2026 08:56:27 +0000

🌱 Green AI: Turning Screen Time into Real-World Adventures

My submission for the Hacktoberfest Open-Source AI Challenge — Week 1: Touch Grass.

🌿 What I Built

I built Green AI, an application designed to encourage people to spend less time staring at screens and more time experiencing the real world.

The idea is simple: instead of spending free time endlessly scrolling, users can generate quests that encourage them to step outside, explore their surroundings, and complete real-world activities.

Green AI makes offline activities feel more engaging by turning them into small adventures.

My goal is to explore how AI can be used not just to help people do things online, but also to encourage them to disconnect and participate in the world around them.

🎬 Demo

I've included a video demonstration of Green AI above.

The demo shows the application and its quest experience.

💻 Code

The project is an opportunity for me to learn about application development, AI integration, debugging, and building practical tools with AI-assisted coding.

🛠️ How I Built It

I developed Green AI with the help of AI-assisted coding and prompts.

During development, I worked on the application structure, frontend-backend integration, and testing.

I also explored the challenges involved in connecting an application to an AI system and verifying that its functionality works as expected.

Building this project helped me understand that developing an application involves more than generating code. It also requires testing, debugging, and verifying that the complete application works.

🌍 Why Does Open Innovation Matter?

Open innovation gives developers opportunities to experiment, learn, collaborate, and build technology that addresses real-world problems.

For Green AI, the broader idea is to explore how AI can encourage healthier digital habits rather than simply keeping people engaged with their screens.

Open-source tools can also help developers understand how software works, adapt it to their needs, and collaborate with others to improve it.

I built this project as a learning experience and as an exploration of how technology can encourage people to spend more time in the real world.

🤖 My Agent Session

I used AI-assisted coding to help develop this project.

🏆 Challenge

This project is my attempt to explore the idea behind Touch Grass: using technology to encourage people to spend more time away from their screens.

I'd love to hear suggestions for improving Green AI and making its quests more useful and engaging.