Skip to content

BeeAI Workshop

Create Tools

Tool Development: Tavily Internet Search + Redis RAG¶

⚠️ Before you begin:
Make sure you have completed the Prework and Setup instructions.
The tools rely on environment configuration, the redis stack to be running, dependencies, and your .env file.

In this section of the workshop, you'll build two essential tools that power your company assistant:

Web Search Tool using a locally run Tavily MCP server
Redis Retriever Tool for enabling Retrieval-Augmented Generation (RAG) from your vector database

Each tool is implemented interactively. You’ll fill in missing logic and test each script on its own before integrating it into your assistant.

Part 1: Tavily MCP Tool¶

Step 1: Verify Your Environment¶

Navigate to the correct directory:

cd beeai-workshop/beeai_fw_tavily_redis

Install dependencies:
```
uv sync
```

Step 2: Locate the Tool File¶

Navigate to the following file:
```
src/tavily_mcp_tool.py
```
Some parts of the implementation are left incomplete for you to fill-in-the-blank and complete the tool. Follow the in-line comments to complete the logic. If you need help, check the solution:
```
src/solutions/tavily_mcp_tool.py
```

Step 3: Run the Tool¶

Run or debug the Tavily tool independently to test its functionality:
```
uv run src/tavily_mcp_tool.py
```

Remember to save the file with your changes before running it!

Part 2: Redis Retriever Tool¶

Step 1: Verify Your Environment¶

If you've just completed Part 1, your environment should still be active. Otherwise:
```
cd beeai-workshop/beeai_fw_tavily_redis
uv sync
```

Step 2: Locate the Tool File¶

Navigate to:
```
src/redis_retriever.py
```
Some parts of the implementation are left incomplete for you to fill-in-the-blank and complete the tool. Follow the in-line comments to complete the logic. If you need help, check the solution:
```
src/solutions/redis_retriever.py
```

Step 3: Run the Tool¶

Run or debug your Redis retriever tool independently:
```
uv run src/redis_retriever.py
```

Remember to save the file with your changes before running it!

This ensures the tool works as expected before integrating it into your BeeAI Agent!