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

What I Built

I built SayLess, a private writing assistant for difficult conversations.

Sometimes you know what you want to say, but finding the right words is hard. SayLess helps by turning a message into three short reply options: Soft, Natural, and Direct. You can also choose what you want to communicate, such as apologizing, explaining, saying no, or setting a boundary.

SayLess is for a friend who finds it hard to write replies to awkward or sensitive messages. It helps them get started while keeping the final choice in their hands. They can copy, edit, or ignore any suggestion.

Demo

Here Youtube link:

The video shows how to enter a message, choose an intent, and get reply suggestions. SayLess runs locally and needs Ollama, so a screen recording is the easiest way to show it working.

Code

Here github repo

rinki538 / Sayless_DEV_Challange1

SayLess is a local AI communication assistant that helps you find the right words for difficult conversations. Paste a message, choose your intent and tone, and get natural reply suggestions powered by an open-source AI model running locally.

SayLess

Find the words. Keep your voice.

SayLess is a private, local-first writing assistant for people who know what they want to say but need help putting it into words. It turns a difficult message into three short reply drafts, or helps you think through the message without drafting a reply.

Your conversations stay on your device when you use Ollama locally.

Built for the Hacktoberfest Weekend Challenge: Build for a Friend.

See it in action

SayLess runs in your browser and connects directly to Ollama on your computer. There is no hosted demo because generation requires a running local Ollama model. Follow Getting started to try it.

Why SayLess?

Awkward or sensitive messages can be hard to answer. Cloud chatbots may require sharing private conversations and can produce replies that do not sound like you. SayLess offers a local alternative: choose what you want to communicate, pick a…

View on GitHub

How I Built It

SayLess is built with React, Vite, and Tailwind CSS. It uses Ollama to run an open-weight AI model on the user’s computer. The default model is qwen3:4b.

The app sends the message and selected options to Ollama, checks the model’s response, and displays the reply drafts. There is also a “Just help me think” option for considering how to respond without writing a reply for you.

Why Does Open Innovation Matter?

Open models let SayLess run on the user’s own computer instead of relying on a hosted AI service. This helps keep sensitive conversations private and gives users more control over the model they use.
Thank You.