# BriefMint

BriefMint turns an independent freelancer's client message into a draft scope, clarification questions, and an editable reply. It runs Gemma 3 through Ollama on the same computer as the app.

Try the [hosted sample-mode demo](https://vibhanshu-s.github.io/briefmint/). It uses a fixed example; live model analysis is available when you run the app locally.

## Run it locally

1. Install [Ollama](https://ollama.com/download) and pull the model once: `ollama pull gemma3:4b`.
2. In this folder, start the app: `python server.py`.
3. Open [http://127.0.0.1:8080](http://127.0.0.1:8080).
4. Use **Try a sample brief** to explore the interface without a model, or paste a real brief and choose **Make the scope** to call your local Gemma model.

The app server binds to loopback only. Brief text is sent only to the Ollama service at `127.0.0.1:11434`; no hosted model API is configured. The app does not store brief text. The sample output is illustrative and is visibly marked as an example. Generated estimates are not rate advice or a binding quote.

## Stack

- Gemma 3 4B open-weight model, served locally by Ollama
- Python standard-library HTTP server
- Plain HTML, CSS, and JavaScript

## Current limits

This is a challenge MVP. The estimate shown for a real model response depends on what the model returns; BriefMint does not know your rates, taxes, or local market. Review every output before sending it to a client. The UI's example analysis is fixed sample content and is not model-generated.