ForgeLink

Cold emails that warm up: from research to relationships.

Table of Contents

1. About The Project
   • Built With
   • Built With
   • Built With
   • Built With
   • Built With
2. Getting Started
3. Contributors
4. Contact
5. Acknowledgments

About The Project

Built with

Demo

Features

• Profile summary:
  BrowserUse → collects text + sources → Gemini extracts 5–10 concise facts (roles, companies, skills, notable work).

• Auto-intent drafting:
  No dropdown needed. The server classifies your prompt into one of:

  • Email (subject + body)
  • Questions (3–6 tailored questions)
  • Message (short DM/intro)
  • Summary (talking points)

• Grounded generation:
  Gemini receives the target’s facts, sources, best URL, and a clipped profile excerpt, plus overlap reasoning with your profile so that drafts stay specific.

• Chat UX built for flow:

  • System intro bubble at the top with profile highlights.
  • Thinking bubble while Gemini drafts.
  • Clean, copy-ready results.

• Targets sidebar (organized outreach):
  Stacked list of people you’ve researched; add from the enter-info page, click to open chat, and delete via a trash icon.

• Simple auth & security:
  JWT signup/login; all targets and actions are bound to the authenticated user (no IDs passed in requests).

• Data model (MongoDB Atlas):

  • Users: username, passwordHash, name, headline, (optional) experiences, interests, tone
  • Targets: ownerUserId, name, headline, facts, sources, profileUrl, profileText, confidence
  • Drafts: ownerUserId, targetId, subject/body or questions

• Dev-friendly setup:
  React + Vite, Express, Nodemon, .env config, robust JSON parsing for LLM outputs, and CORS enabled for local dev.

Color codes

Project Structure

FORGELINK/
├── node_modules/
├── python/.venv
├── public/
│   ├── main.jsx
│   ├── styles.css
│   └── assets/
│       ├── icon.png
├── scripts/
│   ├── collect_profile.py
│   └── seed_user.js
├── server/
│   ├── middleware/
│   │   └── requireAuth.js
│   ├── models/
│   │   └── User.js
│   ├── routes/
│   │   ├── auth.js
│   │   └── targets.js
│   ├── services/
│   │   └── ai.js
│   └── server.js
├── src/
│   ├── components/
│   │   ├── Button.jsx
│   │   ├── Message.jsx
│   │   └── Sidebar.js
│   ├── pages/
│   │   ├── Chat.jsx
│   │   ├── EnterInfo.jsx
│   │   ├── Login.jsx
│   │   └── SignUp.jsx
│   ├── styles/
│   │   ├── Chat.css
│   │   ├── EnterInfo.css
│   │   ├── Login.css
│   │   ├── Sidebar.css
│   │   └── SignUp.css
│   ├── utils/
│   │   ├── api.js
│   │   ├── auth.js
│   │   └── targets.js
│   ├── App.jsx
│   ├── App_test.jsx
│   ├── index.html
│   └── index.jsx
├── .env
├── .gitignore
├── nodemon.json
├── package-lock.json
├── package.json
├── README.md
└── vite.config.js

Getting Started

1. Create a MongoDB atlas account here

2. Download dependencies in npm:

npm init -y
npm i react react-dom react-router-dom cors express mongoose @google/generative-ai bcryptjs jsonwebtoken
npm i -D vite @vitejs/plugin-react concurrently

3. Create virtual env in root directory (ForgeLink)

python3 -m venv python/.venv
source python/.venv/bin/activate

4. Download dependencies with pip:

pip install browser-use python-dotenv playwright
python -m playwright install --with-deps chromium

5. In .env:

PORT=3000
GOOGLE_API_KEY=""
MONGODB_URI=""
NODE_ENV=development
PYTHON_BIN=python/.venv/bin/python
JWT_SECRET=
VITE_API_BASE=

6. Kickstart ForgeLink

npm run dev

Then open http://localhost:5173/!

API will be hosted at port 3000

Contributors

Team member	Roles
Vienna	BrowserUse, Gemini, linking backend and frontend, authentication, Login/Sign up logic, chat page, sidebar
Angela	Designs and aesthetics for pages, login and sign up pages, EnterInfo page

Contact

Vienna Zhao - GitHub - LinkedIn - [email protected]

Angela Ho - GitHub - LinkedIn - [email protected]

Project Link: https://github.com/Angelaho1128/ForgeLink

Acknowledgements

Flaticon

Readme Template 1 - othneildrew

Readme Template 2 - Louis3797