> ## Documentation Index
> Fetch the complete documentation index at: https://agno-v2-studio-tools-doc.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# RAG Agent

## Code

```python cookbook/11_models/ibm/watsonx/knowledge.py theme={null}
from agno.agent import Agent
from agno.knowledge.knowledge import Knowledge
from agno.models.ibm import WatsonX
from agno.vectordb.pgvector import PgVector

db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai"

knowledge_base = Knowledge(
    vector_db=PgVector(table_name="recipes", db_url=db_url),
)
knowledge_base.insert(
  url="https://agno-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"
)

agent = Agent(
    model=WatsonX(id="ibm/granite-20b-code-instruct"),
    knowledge=knowledge_base,
    )
agent.print_response("How to make Thai curry?", markdown=True)
```

## Usage

<Steps>
  <Snippet file="create-venv-step.mdx" />

  <Step title="Set your API key">
    ```bash theme={null}
    export IBM_WATSONX_API_KEY=xxx
    export IBM_WATSONX_PROJECT_ID=xxx
    ```
  </Step>

  <Step title="Install dependencies">
    ```bash theme={null}
    uv pip install -U ibm-watsonx-ai sqlalchemy pgvector psycopg pypdf openai agno
    ```
  </Step>

  <Step title="Set up PostgreSQL with pgvector">
    You need a PostgreSQL database with the pgvector extension installed. Adjust the `db_url` in the code to match your database configuration.
  </Step>

  <Step title="Run Agent">
    ```bash theme={null}
    python cookbook/11_models/ibm/watsonx/knowledge.py
    ```
  </Step>

  <Step title="For subsequent runs">
    After the first run, comment out the `knowledge_base.load(recreate=True)` line to avoid reloading the PDF.
  </Step>
</Steps>

This example shows how to integrate a knowledge base with IBM WatsonX. It loads a PDF from a URL, processes it into a vector database (PostgreSQL with pgvector in this case), and then creates an agent that can query this knowledge base.

Note: You need to install several packages (`pgvector`, `pypdf`, etc.) and have a PostgreSQL database with the pgvector extension available.