🧠 CogniFillz

Smart job application autofill with local AI - by Cranium Inc.

CogniFillz is a privacy-focused, AI-powered job application automation tool that works across desktop and mobile. Unlike Simplify, it uses local LLMs for resume analysis and works on mobile devices through a custom browser.

🌐 Browser Support

Browser	Status	Installation	Notes
Zen Browser	✅ Full Support	Load .xpi	Recommended! Best privacy
Firefox	✅ Full Support	Load .xpi or temporary	Same as Zen
Chrome	✅ Full Support	Load from dist/chrome/	Tested & working
Edge	✅ Full Support	Load from dist/chrome/	Chromium-based
Brave	✅ Full Support	Load from dist/chrome/	Chromium-based

👉 For Zen Browser users: See ZEN_BROWSER_SETUP.md for complete installation guide!

🎯 Features

• Smart Autofill: Automatically detect and fill job application forms
• Cross-Browser: Works on Zen, Firefox, Chrome, Edge, Brave
• AI Resume Scoring: Get a match score for any job posting using your local LLM
• Resume Tailoring: Customize your resume for specific jobs with AI assistance
• Mobile Support: Coming in Phase 3 - custom browser app for mobile
• Privacy First: All AI processing happens locally on your machine
• Free to Use: No subscription fees, works entirely offline

🏗️ Project Structure

cognifillz/
├── apps/
│   ├── extension/      # Chrome Extension (Manifest V3)
│   ├── server/         # FastAPI Backend for LLM integration
│   └── mobile/         # React Native mobile app (Phase 3)
├── packages/
│   ├── shared/         # Shared TypeScript types
│   └── database/       # Supabase client wrapper
└── supabase-schema.sql # Database schema

🚀 Sprint 1: Getting Started (This Week)

Prerequisites

1. Node.js (v18+) and npm
2. Python (v3.9+) and pip
3. LM Studio - Download here
4. Browser: Zen Browser (recommended), Firefox, Chrome, or Edge

👉 Using Zen Browser? Skip to ZEN_BROWSER_SETUP.md for Zen-specific instructions!

Step 1: Set Up LM Studio

1. Download and install LM Studio
2. Download a model (recommended: llama-3.2-3b-instruct for speed or mistral-7b-instruct for quality)
3. Start the local server:
   • In LM Studio, go to "Local Server" tab
   • Click "Start Server"
   • Default endpoint: http://localhost:1234/v1
   • Keep LM Studio running while using CogniFillz

Step 2: Install Extension Dependencies

cd apps/extension
npm install

Step 3: Build the Extension

npm run build

This creates a dist/ folder with your compiled extension.

Step 4: Load Extension in Chrome

1. Open Chrome and go to chrome://extensions/
2. Enable "Developer mode" (top right)
3. Click "Load unpacked"
4. Select the apps/extension/dist folder
5. The CogniFillz icon should appear in your extensions toolbar

Step 5: Create Your Profile

1. Click the CogniFillz icon in Chrome
2. Fill in your information (name, email, LinkedIn, etc.)
3. Click "Save Profile"

Step 6: Test Autofill

1. Navigate to a job application (try LinkedIn Easy Apply or Indeed)
2. Open the CogniFillz popup
3. You should see detected fields listed
4. Click "🚀 Autofill Application"
5. Watch your info get filled in automatically!

🧪 Testing the Extension

Test Sites

Start with these easier sites:

• LinkedIn Easy Apply - Simple forms, good for testing
• Indeed - Straightforward forms
• Greenhouse.io - Common ATS platform

More complex (Phase 2):

• Workday - Uses iframes, more challenging
• Lever - Custom components

What You Should See

1. When you visit a job application page, open DevTools (F12)
2. In the Console, you should see:

   🧠 CogniFillz: Content script loaded
   🧠 CogniFillz: Detected X fields:
     - firstName (90% confidence): First Name
     - email (100% confidence): Email Address
     ...
   

3. Fields should be highlighted with a green outline and labeled
4. The popup should show the count of detected fields

🔧 Development Mode

For faster iteration during development:

# In apps/extension
npm run dev

This watches for file changes and rebuilds automatically. You'll need to click the refresh icon in chrome://extensions/ after each build.

🧠 Phase 2: Set Up the AI Backend (Next Week)

Step 1: Install Python Dependencies

cd apps/server
pip install -r requirements.txt

Step 2: Configure Environment

cp .env.example .env

Edit .env if needed (defaults work for LM Studio).

Step 3: Start the Server

python main.py
# or
uvicorn main:app --reload

Server runs on http://localhost:8000

Step 4: Test the API

Visit http://localhost:8000/docs to see the interactive API documentation.

Test health check:

curl http://localhost:8000/health

🗄️ Phase 2: Set Up Supabase (Optional but Recommended)

Step 1: Create Supabase Project

1. Go to supabase.com
2. Create a free account
3. Create a new project
4. Wait for it to provision (~2 minutes)

Step 2: Run the Schema

1. Go to the SQL Editor in your Supabase dashboard
2. Copy the contents of supabase-schema.sql
3. Paste and run it
4. You should see tables created: profiles, applications, custom_field_mappings

Step 3: Get Your Credentials

1. Go to Settings > API
2. Copy your:
   • Project URL
   • Anon/Public key

Step 4: Configure Extension

Add to apps/extension/src/config.ts:

export const SUPABASE_URL = 'your-project-url';
export const SUPABASE_ANON_KEY = 'your-anon-key';

🎓 Learning Resources

Since you're between beginner and intermediate, here are key concepts:

Chrome Extension Architecture

• manifest.json: Configuration and permissions
• content-script.ts: Runs in the context of web pages
• background.ts: Service worker, handles messaging
• popup: Your UI (React)

Communication flow:

Popup → Background → Content Script → Webpage

Key APIs Used

• chrome.runtime.sendMessage(): Send messages between components
• chrome.storage.local: Store data locally
• chrome.tabs.query(): Get info about tabs
• document.querySelector(): Find elements on page

LLM Integration

The FastAPI server acts as a bridge:

Extension → FastAPI → LM Studio → FastAPI → Extension

This keeps everything local and private.

🐛 Troubleshooting

Fields Not Detected

1. Open DevTools Console - look for CogniFillz logs
2. Check if content script loaded (you should see the 🧠 emoji)
3. Try refreshing the page
4. Some sites load fields dynamically - wait a few seconds

Autofill Not Working

1. Make sure you've created a profile
2. Check that detected fields match your profile data
3. Some sites prevent programmatic filling - this is normal
4. Try manually clicking a field first, then autofill

LLM Not Responding

1. Make sure LM Studio is running
2. Check that local server is started (green indicator)
3. Verify endpoint: http://localhost:1234/v1
4. Test with: curl http://localhost:1234/v1/models

Extension Not Loading

1. Check for errors in chrome://extensions/
2. Make sure you ran npm run build
3. Try removing and re-adding the extension
4. Check that dist/manifest.json exists

📊 Current Status

✅ Completed (Sprint 1)

• Project structure setup
• Chrome extension boilerplate
• Field detection algorithm
• Basic autofill functionality
• React popup UI
• FastAPI server with LLM integration
• Supabase schema

🚧 In Progress (Sprint 2)

• Resume scoring endpoint
• Job description scraping
• Supabase integration in extension
• Field highlighting improvements

📅 Planned (Sprint 3+)

• Resume tailoring with AI
• React Native mobile app
• Advanced ATS support (Workday iframes)
• Heuristic learning for unknown fields
• Job tracking dashboard

💡 Next Steps

1. Test the Extension: Try it on 3-5 different job sites
2. Document Issues: Note which sites work well and which don't
3. Add Your Resume: Expand the profile form to include work experience
4. Test AI Scoring: Get the FastAPI server running and try job analysis

🤝 Contributing

This is a personal project, but feel free to:

• Report bugs
• Suggest features
• Share which job sites work/don't work

📝 License

MIT - Build whatever you want with this!

🎯 Vision

CogniFillz aims to be the most privacy-focused, powerful job application tool. Key differentiators:

1. Local-First: Your data never leaves your machine
2. Mobile Support: Works on phones (coming Phase 3)
3. Free Forever: No subscriptions, no data selling
4. Open Architecture: Use any LLM (Ollama, LM Studio, etc.)

---

Built with 🧠 by Cranium Inc.