HubSpot RAG Assistant

This project is a Retrieval-Augmented Generation (RAG) assistant for HubSpot documentation. It uses OpenAI to generate embeddings, Pinecone to store and search relevant document chunks, and FastAPI to serve a query endpoint.

---

🧩 Features

• Embeds HubSpot documentation from llms-full.txt
• Generates OpenAI embeddings and stores them in Pinecone
• Provides a FastAPI endpoint to query the documents using semantic search
• Optional: Streamlit interface for easy testing

---

🗂️ Project Structure

hubspot-rag-app/
├── llms-full.txt              # Raw documentation source
├── embed.py                   # Script to chunk and embed data into Pinecone
├── query.py                   # Command-line interface for querying the RAG system
├── main.py                    # FastAPI backend for querying and answering
├── streamlit_app.py           # Optional Streamlit UI for local testing
├── .env                       # API keys and config
├── requirements.txt           # Dependencies

---

⚙️ Environment Setup

1. Clone the repository:

git clone <repository-url>
cd hubspot-rag-assistant

2. Create a .env file with your API keys:

OPENAI_API_KEY=your-openai-api-key
PINECONE_API_KEY=your-pinecone-api-key
PINECONE_ENV=your-pinecone-environment

3. Install dependencies:

pip install -r requirements.txt

---

🔧 Populate Pinecone

To generate and store embeddings from the HubSpot documentation:

python embed.py

This script:

• Loads llms-full.txt
• Splits it into chunks
• Generates OpenAI embeddings
• Stores them in Pinecone with IDs

---

🚀 Running the Application

Start the FastAPI Backend

uvicorn main:app --reload

The API will be available at http://localhost:8000

Start the Streamlit Interface

In a new terminal:

streamlit run streamlit_app.py

The Streamlit interface will be available at http://localhost:8501

---

🧪 Test with Postman

Endpoint

POST http://localhost:8000/ask

Request Body

{
  "question": "What are developer test accounts in HubSpot?"
}

Response

{
  "question": "What are developer test accounts in HubSpot?",
  "answer": "Developer test accounts are free HubSpot environments that allow you to test apps and integrations...",
  "sources": [
    "Developer test accounts will expire after 90 days if no API calls...",
    "You can create up to 10 test accounts per developer account..."
  ]
}

---

💡 Optional: Streamlit UI

To test locally with a simple web interface:

streamlit run streamlit_app.py

You'll get a local UI where you can enter questions and see answers and source snippets like in the example below:

🧐 Technologies Used

• OpenAI
• Pinecone
• LangChain
• FastAPI
• Streamlit (optional)

---

📌 Roadmap

• Add auth and usage limits
• Deploy to cloud (e.g., Render/Fly.io)
• Enable document updates/re-indexing

---

🙌 Credits

• HubSpot Developer Docs: https://developers.hubspot.com/docs
• Built with OpenAI, Pinecone, FastAPI

💻 Command-Line Interface

For quick testing or integration with other tools, you can use the command-line interface:

python3 query.py

The CLI will:

1. Prompt you to enter a question
2. Show the most relevant documentation chunks
3. Optionally generate a GPT-4 answer based on the chunks

Example usage:

Ask a question about HubSpot development: What are developer test accounts?
🔍 Top Chunks:
[Shows relevant documentation chunks]

🧠 Use GPT-4 to generate answer from these chunks? (y/n): y
✅ Answer:
[Shows generated answer]

🐛 Troubleshooting

1. If you see "command not found" errors:
   • Verify all dependencies are installed (pip install -r requirements.txt)