Prompt Mirror

Visualize your thinking patterns & watch your mind evolve

Live Demo → · Report Bug · Request Feature

---

What is Prompt Mirror?

Prompt Mirror is an open-source AI usage analytics platform that passively captures the prompts you send across major AI tools, then surfaces rich insights about your thinking patterns, topic clusters, and cognitive habits — all from your own data.

"You can't improve what you can't measure."

Most people have no idea how they actually use AI. Prompt Mirror acts as a mirror — giving you an honest, data-driven reflection of your AI interactions.

---

✨ Features

Category	Highlights
Capture	Auto-captures prompts from ChatGPT, Claude, Gemini, Trae, OpenAI Playground
Analytics	Topic modelling, word clouds, length trends, hourly heatmaps, source breakdown
Insights	AI-powered thinking style analysis; tracks how your prompts evolve over time
Privacy	All data stays in your own account (or fully local if you skip cloud sync)
Sync	Optional cloud sync across devices via GitHub or Google OAuth
i18n	English / 中文 UI toggle built-in

---

Architecture

┌────────────────────────────────────────────────────────────────┐
│                        User's Browser                          │
│                                                                │
│  ┌──────────────────────┐      ┌──────────────────────────┐   │
│  │   Chrome Extension   │      │  Streamlit Dashboard     │   │
│  │   (Manifest V3)      │      │  prompt_mirror.py        │   │
│  │                      │      │                          │   │
│  │  content.js injects  │      │  • Plotly charts         │   │
│  │  into AI chat pages  │      │  • Word clouds           │   │
│  │  → captures prompts  │      │  • Heatmaps              │   │
│  │  → stores locally    │      │  • Topic analysis        │   │
│  └──────────┬───────────┘      └────────────┬─────────────┘   │
│             │  Export JSON / Cloud Sync      │ REST API        │
└─────────────┼──────────────────────────────-┼─────────────────┘
              │                               │
              ▼                               ▼
     ┌────────────────────────────────────────────────┐
     │               FastAPI Backend                  │
     │          (prompt-mirror-api)                   │
     │                                                │
     │   /auth/github   /auth/google   (OAuth 2.0)    │
     │   /api/prompts   /api/stats     (JWT Auth)     │
     └────────────────────────┬───────────────────────┘
                              │ SQLAlchemy ORM
                              ▼
                   ┌──────────────────┐
                   │   PostgreSQL DB  │
                   │ (prompt-mirror-  │
                   │      db)         │
                   └──────────────────┘

---

Supported AI Platforms

Platform	Status
ChatGPT (chatgpt.com)	✅ Supported
Claude (claude.ai)	✅ Supported
Google Gemini	✅ Supported
Trae AI	✅ Supported
OpenAI Playground	✅ Supported

---

Quick Start

Option 1 — Local (No backend needed)

# 1. Clone the repo
git clone https://github.com/MaxonT/prompt-mirror.git
cd prompt-mirror

# 2. Install dependencies
pip install -r requirements.txt

# 3. Run the dashboard
streamlit run prompt_mirror.py

Open http://localhost:8501, then load a JSON export from the Chrome extension.

---

Option 2 — Full Stack (with cloud sync)

Prerequisites

• Python 3.11+
• PostgreSQL (or use Render free tier)
• GitHub OAuth App and/or Google OAuth credentials

1. Clone & configure environment

git clone https://github.com/MaxonT/prompt-mirror.git
cd prompt-mirror
cp .env.example .env

Edit .env and fill in your credentials:

DATABASE_URL=postgresql://user:password@localhost:5432/prompt_mirror
JWT_SECRET=your-long-random-secret

# GitHub OAuth — https://github.com/settings/applications/new
GITHUB_CLIENT_ID=
GITHUB_CLIENT_SECRET=

# Google OAuth — https://console.cloud.google.com/apis/credentials
GOOGLE_CLIENT_ID=
GOOGLE_CLIENT_SECRET=

API_BASE_URL=http://localhost:8000
FRONTEND_URL=http://localhost:8501

2. Start the backend

pip install -r backend/requirements.txt
uvicorn backend.app:app --reload --port 8000

API docs available at http://localhost:8000/docs

3. Start the frontend

pip install -r requirements.txt
streamlit run prompt_mirror.py

---

Chrome Extension Setup

1. Open Chrome → chrome://extensions/
2. Enable Developer mode (top-right toggle)
3. Click Load unpacked → select the chrome-extension/ folder
4. Pin the extension to your toolbar
5. Visit any supported AI platform — prompts are captured automatically

To export your data: Click the extension icon → Export JSON → drag the file into the Prompt Mirror dashboard sidebar.

---

One-Click Deploy to Render

The project ships with a render.yaml Blueprint for zero-config deployment.

After deploying, set these environment variables in your Render dashboard:

Variable	Description
GITHUB_CLIENT_ID	From GitHub OAuth App settings
GITHUB_CLIENT_SECRET	From GitHub OAuth App settings
GOOGLE_CLIENT_ID	From Google Cloud Console
GOOGLE_CLIENT_SECRET	From Google Cloud Console

DATABASE_URL and JWT_SECRET are provisioned and injected automatically by the Blueprint.

OAuth Callback URLs to configure:

Provider	Callback URL
GitHub	https://prompt-mirror-api.onrender.com/auth/github/callback
Google	https://prompt-mirror-api.onrender.com/auth/google/callback

---

Tech Stack

Layer	Technology
Frontend	Streamlit, Plotly, pandas, WordCloud, jieba, Matplotlib
Backend	FastAPI, SQLAlchemy, Pydantic, python-jose (JWT)
Database	PostgreSQL
Extension	Chrome Manifest V3, Vanilla JS
Auth	GitHub OAuth 2.0, Google OAuth 2.0
Deployment	Render (Blueprint), Docker

---

Project Structure

prompt-mirror/
├── prompt_mirror.py        # Streamlit analytics dashboard
├── requirements.txt        # Frontend dependencies
├── render.yaml             # Render deployment Blueprint
├── Dockerfile              # Frontend container
├── logo.png                # Brand logo / favicon
│
├── backend/                # FastAPI backend
│   ├── app.py              # Main application, routes
│   ├── auth.py             # OAuth + JWT logic
│   ├── database.py         # SQLAlchemy engine & session
│   ├── models.py           # ORM models (User, Prompt)
│   ├── schemas.py          # Pydantic schemas
│   └── requirements.txt    # Backend dependencies
│
├── chrome-extension/       # Browser extension
│   ├── manifest.json       # Extension manifest (MV3)
│   ├── content.js          # Prompt capture logic
│   ├── background.js       # Service worker
│   ├── popup.html/js/css   # Extension UI
│   └── icons/              # Extension icons
│
└── .streamlit/
    └── config.toml         # Streamlit server config

---

Environment Variables Reference

Variable	Required	Description
DATABASE_URL	✅	PostgreSQL connection string
JWT_SECRET	✅	Secret key for signing JWTs
GITHUB_CLIENT_ID	Optional	GitHub OAuth App client ID
GITHUB_CLIENT_SECRET	Optional	GitHub OAuth App client secret
GOOGLE_CLIENT_ID	Optional	Google OAuth client ID
GOOGLE_CLIENT_SECRET	Optional	Google OAuth client secret
API_BASE_URL	✅	Public URL of the backend API
FRONTEND_URL	✅	Public URL of the Streamlit frontend

---

Contributing

Contributions are welcome! Here's how to get started:

# 1. Fork and clone
git clone https://github.com/YOUR_USERNAME/prompt-mirror.git

# 2. Create a feature branch
git checkout -b feat/your-feature-name

# 3. Make your changes, then commit
git commit -m "feat: add awesome feature"

# 4. Push and open a Pull Request
git push origin feat/your-feature-name

Please follow Conventional Commits for commit messages.

---

Roadmap

• Firefox extension support
• CSV / Markdown export
• Semantic clustering with embeddings
• Weekly digest email
• Self-hostable Docker Compose stack
• Per-platform breakdown in timeline view

---

License

Distributed under the MIT License. See LICENSE for full details.

---

Acknowledgements

Built with Streamlit, FastAPI, and a lot of curiosity about how humans think with AI.

---

_{Made with ♥ · Live Demo}