DEV Community: Hitesh Kumar

CartSpark — Autonomous E-Commerce Listing Studio Powered by Google Gemma 2

Hitesh Kumar — Mon, 26 Oct 2026 05:55:31 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady

What I Built

CartSpark is an autonomous e-commerce listing generator, unit economics price optimizer, and customer review copilot designed specifically for marketplace sellers and independent creators (Etsy, Amazon, eBay, and Shopify).

The Problem It Solves

Ask any independent maker, vintage curator, or small e-commerce brand owner what consumes their most painful hours, and the answer is always the same: writing listings, calculating margins, and dealing with reviews.

Writing a high-ranking product title within marketplace character caps (e.g. Etsy's 140-character limit).
Brainstorming 13 exact marketplace search tags with high search volume.
Crafting 5 benefit-driven bullet points that convert casual scrollers into paying customers.
Calculating break-even charm pricing ($29.95 vs $34.00) after payment and marketplace fees eat away profits.
Patiently drafting de-escalation responses to 1-star or 2-star customer feedback. A seller with 15 new weekly inventory items spends 3 to 4 hours doing this manually every week. ### Who Would Pay For It? Side-hustlers and boutique merchants happily pay $15 to $29/month for listing tools. However, conventional cloud AI tools charge per-generation token fees or lock users into costly monthly subscriptions with restrictive rate limits. CartSpark changes the equation by running on Google's open-weight Gemma 2 model:
$0 Marginal API Cost: Unlimited listings without per-token subscription bills.
100% Data Privacy: Unreleased product concepts, proprietary materials, and cost margins never leave the seller's computer.
Offline Capable: Runs anywhere—in craft workshops, flea markets, or basement warehouses without Wi-Fi.
Live Application: https://cartspark-ai.onrender.com (Hosted on Render)
Local Live Studio: http://127.0.0.1:8000
UI & Experience: Crafted with an obsidian dark theme, clean typography, responsive layout, and 100% clean vector SVG icons with zero generic emojis.

Code

The complete source code is open-source on GitHub:

hitesh-kumar123 / cartspark-ai

CartSpark ⚡

Autonomous E-Commerce Listing Studio Powered by Google Gemma 2

Hacktoberfest Open-Source AI Challenge: Week 2 — Side Hustle
Targeted Prize Categories: Overall Winner, Best Use of Gemma, Best Use of Render

📌 Overview

CartSpark is an AI-powered e-commerce studio built for independent artisans, Etsy sellers, Amazon FBA merchants, and Shopify store owners.

Writing compelling, SEO-ranked product titles, 13 marketplace search tags, benefit-driven bullets, and handling customer feedback can take 15–30 minutes per product. For a seller adding 10–20 products a day, this is tedious, repetitive work.

While cloud AI APIs (like OpenAI or Anthropic) can assist, they introduce a crippling problem for small businesses: ongoing token bills and data privacy risks. Every generation costs money, and uploading unreleased designs or cost margins leaks sensitive trade secrets.

CartSpark solves this at the root: Powered by Google's open-weight Gemma 2 model, it runs on-device or on self-hosted…

View on GitHub

Repository Link: https://github.com/hitesh-kumar123/cartspark-ai

Key Architecture:

app.py: High-performance FastAPI server serving API endpoints and static SPA files.
gemma_engine.py: Google Gemma 2 inference pipeline with local Ollama support (gemma2:2b / gemma2:9b) and an edge reasoning engine.
static/index.html: Studio workspace layout with multi-marketplace selectors and unit economics widgets.
static/css/style.css: Bespoke design system (Inter + JetBrains Mono typography, dark mode glassmorphism).
static/js/app.js: Reactive frontend controller handling state, live character counters, 1-click clipboard actions, and local catalog persistence.
render.yaml: Production blueprint for automated deployment on Render.

How I Built It

CartSpark is engineered around Google's Gemma 2 open-weight model:

1. The Open-Source AI Core: Google Gemma 2

I chose Gemma 2 (2B) for its exceptional reasoning-to-parameter ratio and fast inference on standard laptop hardware.

Domain-Specific Prompt Engineering: The engine structures output to fit strict marketplace constraints (e.g. emotional storytelling and craft authenticity for Etsy vs. punchy specifications and benefit-driven bullets for Amazon).
Inference Flexibility: CartSpark connects to a local Ollama instance running gemma2:2b or gemma2:9b. On edge environments, it automatically transitions to a built-in reasoning engine, ensuring zero downtime. ### 2. Multi-Marketplace Intelligence Marketplaces are not one-size-fits-all:
Etsy Mode: Enforces 13 specific search tags under 20 characters and emphasizes craft heritage.
Amazon Mode: Generates 5 benefit-driven feature bullets following the high-converting (Feature: Benefit) format.
Shopify & eBay: Generates clean titles with clear specification and condition metadata. ### 3. Dynamic Unit Economics & Charm Pricing A side hustle only succeeds if the numbers work. CartSpark includes a dedicated pricing module that calculates: $$\text{Optimal Retail Price} = \frac{\text{COGS} + \text{Shipping}}{1 - \text{Marketplace Fee \%} - \text{Target Margin \%}}$$ It automatically formats this into psychological charm pricing (e.g. $34.95) and delivers positioning advice explaining how to justify the price to buyers. ### 4. Privacy-First Local Catalog Every generated listing can be saved with one click to a Local Persistent Catalog (stored via browser LocalStorage). Proprietary designs, upcoming collection names, and wholesale costs never leave the user's laptop. Sellers can batch-export their entire catalog as a clean CSV ready for bulk upload. --- ## Why Does Open Innovation Matter? This project showcases why open-source AI is the foundational enabler for independent micro-entrepreneurs:
The Reality of Side Hustle Unit Economics ($0 API Bills): If an Etsy seller making $500/month has to pay $30 to $50/month in cloud LLM API tokens, that represents 10% of their net take-home pay. By running Gemma 2 locally, the marginal cost per listing is literally zero dollars. The seller can generate 50 variations of a title without worrying about their credit card getting billed.
Complete Confidentiality & Trade Secret Protection: Craftspeople and niche creators often spend months developing proprietary techniques, unique resin formulas, or unreleased holiday collections. Uploading these drafts to closed cloud APIs risks having proprietary product concepts scraped or ingested into public training sets. With on-device open models, data privacy is absolute.
Offline Reliability in the Field: Makers often work in basements, rural craft sheds, or busy flea markets with patchy Wi-Fi. CartSpark running on Gemma 2 operates 100% offline with zero external network dependencies.
Freedom from Vendor Lock-In: Closed API providers regularly change rate limits, deprecate endpoints, or increase token pricing. Open weights belong to the community forever. ---

My First User

To validate CartSpark with a real merchant, I tested the tool with Elena, an independent ceramicist who runs a pottery shop selling handmade mugs and pour-over drippers:

"Writing Etsy tags is my least favorite part of the entire week. I usually just guess five words and give up. With CartSpark, I pasted my clay type and dimensions, and in two seconds I had 13 comma-separated tags that actually made sense, plus a product description that sounded way more professional than what I usually write.

The price calculator was an eye-opener too—I realized I was undercharging by almost $6 on my drippers after Etsy's transaction and payment processing cuts. I'd easily pay $20 for this if it saves me an hour every restock day."

Elena’s feedback directly inspired the 1-click Copy All Tags feature and the CSV Catalog Export so she could upload batches directly to Etsy.

Prize Categories

I am officially entering CartSpark into the following Hacktoberfest Week 2 categories:

Overall Hacktoberfest Week 2 Winner (Side Hustle Track)

Best Use of Gemma ($200 USD Track): Google’s open-weight Gemma 2 is the core AI model powering CartSpark's listing generator, SEO tag matrix, and customer review resolution copilot.

Best Use of Render ($200 USD Track): The application is fully containerized and configured for instantaneous deployment on Render via the included render.yaml specification.

CartSpark — Autonomous E-Commerce Listing Studio Powered by Google Gemma 2

Hitesh Kumar — Sat, 17 Oct 2026 12:46:05 +0000

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

What I Built

The Problem It Solves

Writing a high-ranking product title within marketplace character caps (e.g. Etsy's 140-character limit).
Brainstorming 13 exact marketplace search tags with high search volume.
Crafting 5 benefit-driven bullet points that convert casual scrollers into paying customers.
Calculating break-even charm pricing ($29.95 vs $34.00) after payment and marketplace fees eat away profits.
Patiently drafting de-escalation responses to 1-star or 2-star customer feedback. A seller with 15 new weekly inventory items spends 3 to 4 hours doing this manually every week. ### Who Would Pay For It? Side-hustlers and boutique merchants happily pay $15 to $29/month for listing tools. However, conventional cloud AI tools charge per-generation token fees or lock users into costly monthly subscriptions with restrictive rate limits. CartSpark changes the equation by running on Google's open-weight Gemma 2 model:
$0 Marginal API Cost: Unlimited listings without per-token subscription bills.
100% Data Privacy: Unreleased product concepts, proprietary materials, and cost margins never leave the seller's computer.
Offline Capable: Runs anywhere—in craft workshops, flea markets, or basement warehouses without Wi-Fi.

Demo

Live Application: https://cartspark-ai.onrender.com (Hosted on Render)
Local Live Studio: http://127.0.0.1:8000

Code

The complete source code is open-source on GitHub:

hitesh-kumar123 / cartspark-ai

CartSpark ⚡

Autonomous E-Commerce Listing Studio Powered by Google Gemma 2

Hacktoberfest Open-Source AI Challenge: Week 2 — Side Hustle
Targeted Prize Categories: Overall Winner, Best Use of Gemma, Best Use of Render

📌 Overview

CartSpark is an AI-powered e-commerce studio built for independent artisans, Etsy sellers, Amazon FBA merchants, and Shopify store owners.

CartSpark solves this at the root: Powered by Google's open-weight Gemma 2 model, it runs on-device or on self-hosted…

View on GitHub

Repository Link: https://github.com/hitesh-kumar123/cartspark-ai

Key Architecture:

app.py: High-performance FastAPI server serving API endpoints and static SPA files.
gemma_engine.py: Google Gemma 2 inference pipeline with local Ollama support (gemma2:2b / gemma2:9b) and an edge reasoning engine.
static/index.html: Studio workspace layout with multi-marketplace selectors and unit economics widgets.
static/css/style.css: Bespoke design system (Inter + JetBrains Mono typography, dark mode glassmorphism).
static/js/app.js: Reactive frontend controller handling state, live character counters, 1-click clipboard actions, and local catalog persistence.
render.yaml: Production blueprint for automated deployment on Render.

How I Built It

CartSpark is engineered around Google's Gemma 2 open-weight model:

1. The Open-Source AI Core: Google Gemma 2

I chose Gemma 2 (2B) for its exceptional reasoning-to-parameter ratio and fast inference on standard laptop hardware.

Domain-Specific Prompt Engineering: The engine structures output to fit strict marketplace constraints (e.g. emotional storytelling and craft authenticity for Etsy vs. punchy specifications and benefit-driven bullets for Amazon).
Inference Flexibility: CartSpark connects to a local Ollama instance running gemma2:2b or gemma2:9b. On edge environments, it automatically transitions to a built-in reasoning engine, ensuring zero downtime.

2. Multi-Marketplace Intelligence

Marketplaces are not one-size-fits-all:

Etsy Mode: Enforces 13 specific search tags under 20 characters and emphasizes craft heritage.
Amazon Mode: Generates 5 benefit-driven feature bullets following the high-converting (Feature: Benefit) format.
Shopify & eBay: Generates clean titles with clear specification and condition metadata.

3. Dynamic Unit Economics & Charm Pricing

A side hustle only succeeds if the numbers work. CartSpark includes a dedicated pricing module that calculates:
$$\text{Optimal Retail Price} = \frac{\text{COGS} + \text{Shipping}}{1 - \text{Marketplace Fee \%} - \text{Target Margin \%}}$$
It automatically formats this into psychological charm pricing (e.g. $34.95) and delivers positioning advice explaining how to justify the price to buyers.

4. Privacy-First Local Catalog

Every generated listing can be saved with one click to a Local Persistent Catalog (stored via browser LocalStorage). Proprietary designs, upcoming collection names, and wholesale costs never leave the user's laptop. Sellers can batch-export their entire catalog as a clean CSV ready for bulk upload.

Why Does Open Innovation Matter?

This project showcases why open-source AI is the foundational enabler for independent micro-entrepreneurs:

The Reality of Side Hustle Unit Economics ($0 API Bills): If an Etsy seller making $500/month has to pay $30 to $50/month in cloud LLM API tokens, that represents 10% of their net take-home pay. By running Gemma 2 locally, the marginal cost per listing is literally zero dollars. The seller can generate 50 variations of a title without worrying about their credit card getting billed.
Complete Confidentiality & Trade Secret Protection: Craftspeople and niche creators often spend months developing proprietary techniques, unique resin formulas, or unreleased holiday collections. Uploading these drafts to closed cloud APIs risks having proprietary product concepts scraped or ingested into public training sets. With on-device open models, data privacy is absolute.
Offline Reliability in the Field: Makers often work in basements, rural craft sheds, or busy flea markets with patchy Wi-Fi. CartSpark running on Gemma 2 operates 100% offline with zero external network dependencies.
Freedom from Vendor Lock-In: Closed API providers regularly change rate limits, deprecate endpoints, or increase token pricing. Open weights belong to the community forever.

My Agent Session

This project was conceived, designed, and built collaboratively using an autonomous AI pair programmer (Google DeepMind Antigravity). The full architectural planning, prompt design, and iterative testing were guided step-by-step to optimize for local Gemma 2 execution, zero-cost unit economics, and clean user experience.

My First User

To validate CartSpark with a real merchant, I tested the tool with Elena, an independent ceramicist who runs a pottery shop selling handmade mugs and pour-over drippers:

"Writing Etsy tags is my least favorite part of the entire week. I usually just guess five words and give up. With CartSpark, I pasted my clay type and dimensions, and in two seconds I had 13 comma-separated tags that actually made sense, plus a product description that sounded way more professional than what I usually write.

The price calculator was an eye-opener too—I realized I was undercharging by almost $6 on my drippers after Etsy's transaction and payment processing cuts. I'd easily pay $20 for this if it saves me an hour every restock day."
Elena’s feedback directly inspired the 1-click Copy All Tags feature and the CSV Catalog Export so she could upload batches directly to Etsy.

Prize Categories

I am officially entering CartSpark into the following Hacktoberfest Week 2 categories:

Overall Hacktoberfest Week 2 Winner (Side Hustle Track)
Best Use of Gemma ($200 USD Track): Google’s open-weight Gemma 2 is the core AI model powering CartSpark's listing generator, SEO tag matrix, and customer review resolution copilot.
Best Use of Render ($200 USD Track): The application is fully containerized and configured for instantaneous deployment on Render via the included render.yaml specification.

Voice2Memory: Never Lose an Important Task in a Voice Note Again

Hitesh Kumar — Mon, 05 Oct 2026 04:22:19 +0000

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend

What I Built

I built Voice2Memory for my friend Alex, a busy founder who records dozens of audio voice notes every week while commuting and moving between meetings.

The Problem

Voice notes are write-only memory. Finding a specific task, deadline, or promise later required Alex to replay minutes of audio at 1x speed. As a result, critical commitments got forgotten, deadlines were missed, and useful ideas remained trapped inside unsearchable audio files.

What It Does & How It Solves It

Voice2Memory automatically converts raw voice notes into structured, searchable personal memories:

Actionable Tasks: Extracts clear to-do items from spoken thoughts.
Dates & Deadlines: Identifies exact meeting times and commitments.
People & Topics: Detects names mentioned and categorizes notes with tags.
Searchable Vault: Lets Alex search across past recordings instantly instead of listening back to audio files.

Demo

Live Application: https://voice2memory.onrender.com
Video Walkthrough: https://drive.google.com/file/d/1dZRGLBKB8y0aG1TXI1AOgyJMCSVXrNZY/view?usp=sharing

Code

hitesh-kumar123 / Voice2Memory

Transform raw voice notes into structured, actionable personal memories with Google Gemma 2 open-weight AI, Faster-Whisper, and MongoDB Atlas.

Voice2Memory

Transform raw voice notes into structured, actionable personal memories using open-weight AI.
Built for the Hacktoberfest 2026 DEV Weekend Challenge: "Build for a Friend".

Overview

The Friend Problem

My friend records dozens of voice notes throughout the week—while commuting, walking between meetings, or capturing sudden startup ideas and daily errands.

A typical spoken recording sounds like this:

"Hey Sarah, I'm heading over to the office now. Please remember to finalize the Q4 investor pitch deck with Alex by Thursday at 4 PM. We also need to schedule the follow-up meeting with Priya for next Monday afternoon. On my way back I have to order a new microphone for our podcast setup and email the contract to Tom."

The problem? Voice notes are write-only memory.

Days later, finding a specific deadline, promise, or action item requires replaying minutes of audio at 1x speed. Crucial commitments get lost, deadlines slip…

View on GitHub

How I Built It

Voice2Memory is built around local open-weight AI models to process voice notes privately and deterministically:

Open-Weight AI Brain (Google Gemma 2): Used Google's gemma2:2b instruction-tuned open-weight model running locally via Ollama. It takes raw transcripts and extracts structured JSON containing executive summaries, action items, dates, people, and topics.
Speech-to-Text (Faster-Whisper): Used faster-whisper (CTranslate2 int8 backend) with Voice Activity Detection (VAD) for fast, offline transcription directly from audio buffers.
Database & Storage (MongoDB Atlas): Used MongoDB Atlas with weighted text search indexes to store and instantly query structured memories by topic, name, or keyword, with live task completion persistence.
Frontend & App Framework: Built with Next.js 16 (App Router), React 19, TypeScript, and Tailwind CSS.

Why Does Open Innovation Matter?

Voice notes are deeply personal—they contain unfiltered thoughts, private work updates, financial decisions, and personal appointments.

Open innovation made three things possible that closed proprietary APIs could not:

Absolute Privacy: By pairing local Faster-Whisper with Google's open-weight Gemma 2, audio files and transcripts never need to leave the user's infrastructure or be fed into third-party AI training pipelines.
Zero Recurring API Costs: Closed APIs charge per-minute for audio transcription and per-token for LLM calls, making continuous daily voice logging expensive. Open-weight models run locally at near-zero incremental cost.
Reliability & Independence: Open-weight models ensure the application will continue running without vendor lock-in, rate limits, or sudden API deprecations.

My Agent Session

Prize Categories

1. Best Use of Gemma ($200)

Model Used: Google Gemma 2 (gemma2:2b open-weight model).
How It Was Used: Powers the core intelligence engine in src/lib/ollama.ts and /api/analyze. It processes raw voice transcripts through custom Gemma instruction templates (<start_of_turn>user ...) and outputs deterministic JSON containing summaries, actionable tasks, temporal dates, people, and topics.

2. Best Use of MongoDB Atlas ($100)

How It Was Used: Serves as the central persistent memory store in src/lib/db.ts. Utilizes weighted compound text search indexes across titles, summaries, transcripts, topics, and people, and supports real-time task checklist synchronization via PATCH /api/memories/[id].

Thanks for checking out Voice2Memory! Feel free to explore the repository, try the live demo, and share any feedback or thoughts in the comments below.

What Open Source Has Taught Me So Far

Hitesh Kumar — Fri, 02 Oct 2026 16:38:32 +0000

What Open Source Has Taught Me So Far

I started contributing to open source because I wanted to learn how real projects are built outside of tutorials.

Over time, I got comfortable working with technologies like React, Node.js, TypeScript, MongoDB, Next.js, and REST APIs. But the biggest learning didn't come from the technologies themselves.

It came from reading someone else's code, understanding an existing codebase, opening issues, reviewing pull requests, fixing bugs, and learning how to communicate with other developers.

Being involved in open source has also taught me that you don't need to know everything before making your first contribution.

You can start small.

Fix a bug.
Improve documentation.
Pick up a beginner-friendly issue.
Help another contributor.
Ask questions.

Every small contribution teaches something.

I'm continuing to explore open source, build projects, and learn from the developer community.

Looking forward to learning more and contributing more during Hacktoberfest 2026.

hacktoberfest #opensource #webdevelopment #javascript #react #nodejs

DEV Community: Hitesh Kumar

CartSpark — Autonomous E-Commerce Listing Studio Powered by Google Gemma 2

What I Built

The Problem It Solves

Code

hitesh-kumar123 / cartspark-ai

CartSpark ⚡

Autonomous E-Commerce Listing Studio Powered by Google Gemma 2

📌 Overview

Key Architecture:

How I Built It

1. The Open-Source AI Core: Google Gemma 2

My First User

Elena’s feedback directly inspired the 1-click Copy All Tags feature and the CSV Catalog Export so she could upload batches directly to Etsy.

Prize Categories

CartSpark — Autonomous E-Commerce Listing Studio Powered by Google Gemma 2

What I Built

The Problem It Solves

Demo

Code

hitesh-kumar123 / cartspark-ai

CartSpark ⚡

Autonomous E-Commerce Listing Studio Powered by Google Gemma 2

📌 Overview

Key Architecture:

How I Built It

1. The Open-Source AI Core: Google Gemma 2

2. Multi-Marketplace Intelligence

3. Dynamic Unit Economics & Charm Pricing

4. Privacy-First Local Catalog

Why Does Open Innovation Matter?

My Agent Session

My First User

Prize Categories

Voice2Memory: Never Lose an Important Task in a Voice Note Again

What I Built

The Problem

What It Does & How It Solves It

Demo

Code

hitesh-kumar123 / Voice2Memory

Transform raw voice notes into structured, actionable personal memories with Google Gemma 2 open-weight AI, Faster-Whisper, and MongoDB Atlas.

Voice2Memory

Overview

The Friend Problem

How I Built It

Why Does Open Innovation Matter?

My Agent Session

Prize Categories

1. Best Use of Gemma ($200)

2. Best Use of MongoDB Atlas ($100)

What Open Source Has Taught Me So Far

What Open Source Has Taught Me So Far

hacktoberfest #opensource #webdevelopment #javascript #react #nodejs