This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady
What I Built
I am building InboxZen AI, an AI-powered email and message productivity assistant inspired by the idea of working smarter, not harder.
Writing replies, drafting emails, and summarizing long messages can take up valuable time. InboxZen AI aims to make these everyday tasks easier with a simple interface and an open-weight AI model.
The project is designed around three core features:
AI Reply Generator: Generate context-aware replies in professional, friendly, formal, polite, or concise tones.
Email Drafter: Turn a short description into an editable email with a subject and body.
Message Summarizer: Extract key points and action items from long messages.
Users can review, edit, and copy the generated content instead of starting every message from scratch.
The goal is not to send emails without human involvement. It is to reduce repetitive writing while keeping the user in control.
Demo
Live application-https://inboxzen-ai.netlify.app/
Code
GitHub repository: https://github.com/priyarani22023-max/InboxZen-AI
The repository includes the frontend, serverless AI functions, deployment configuration, and setup instructions.
How I Built It
I used Google AI Studio to help build and refine the application, React and Vite for the frontend, and Netlify for hosting and serverless backend functions.
The AI integration uses Google's hosted API to access the open-weight Gemma 4 26B A4B Instruct model (gemma-4-26b-it). The application sends requests through a server-side Netlify Function so that the API key is not exposed in browser code.
The main request flow is:
The user enters a message and selects an AI task.
The frontend sends the request to a Netlify Function.
The function validates the input and calls Gemma.
The generated result is returned to the interface for review and editing.
The model is configurable through an environment variable, allowing the inference configuration to be changed without embedding model settings in the frontend.
Why Does Open Innovation Matter?
Open innovation makes it possible for developers to build on open-weight models, experiment with their behavior, and explore different deployment approaches.
For InboxZen AI, Gemma provides an open-weight model that can power everyday text-generation tasks. This creates opportunities to study the model, experiment with prompts, and potentially move to a self-hosted inference setup in the future.
For this initial version, inference is hosted through an API rather than running locally. That means the application still depends on an internet connection and an external inference service. Text submitted for generation is sent to that service, so users should avoid entering sensitive information.
A future local-inference version could offer more control over where personal messages are processed. Open-weight AI gives developers room to explore that possibility instead of being locked into one model or deployment approach.
Top comments (0)