This guided sample is the quickest way to get started with Griptape Cloud.
You'll learn how to create a data source with live news data from the web, a fully-managed vector index knowledge base to serve up your data, and chat assistant that can query and retrieve your data. Use it to answer questions, summarize information, or explain the meaning of life.
Follow these steps to create a sample chat assistant connected to a knowledge base with information from your data source.
Step 1: Create a data source
Log in to Griptape Cloud.
Navigate to the Home screen.
Click
Get Started
in the section titledData retrieval chat: Daily news
.A guide banner will appear at the top of the page to provide step-by-step instructions.
Click
Next
in the banner to navigate to the data source creation form.We've pre-populated the form with URLs to a few popular daily news websites. Feel free to add or remove websites.
(Optional) change the name or description of the data source.
Click
Create
to submit the form.You will be directed to the data source detail page.
What's happening
Creating the data source will automatically initiate a process of extracting, cleaning, and transforming the web page content. This process is known as a data job. While the job is in progress, you can observe the job status and other details.
For this sample, the job should take around a minute. Please be patient; this is longest step!
Step 2: Create a knowledge base containing your data
Once the data job has completed, click
Next
in the banner to navigate to the knowledge base creation form.We've pre-populated the form with the data source you just created. Feel free to select another data source if you already have one.
(Optional) change the name or description of the knowledge base.
Click
Create
to submit the form.You will be directed to the knowledge base detail page.
What's happening
Creating the knowledge base will automatically load your data source content into a vector database index that is optimized for LLMs to query and retrieve information. This process is known as a knowledge base job. While the job is in progress, you can observe the job status and other details.
For this sample, the job should take around half a minute.
Step 3: Create an assistant with access to your knowledge base
Once the knowledge base job has completed, click
Next
in the banner to navigate to the assistant creation screen.We've pre-populated the form with the knowledge base you just created.
(Optional) change the name or description of the knowledge base.
Click Create to submit the form.
You will be directed to the chat interface for your new assistant.
Type a message to start chatting! Try asking, "What's some good news?"
By default, your chat will be written to a new conversation thread. To start a new thread or resume a previous one, use the chat window controls.
Click
Done
in the banner to close the guide.
Congratulations! You've just finished creating a data retrieval assistant that can chat about current events.
What's next
Now that you're familiar with data sources, knowledge bases, and assistants, you can experiment with your assistant or try additional samples. Here are some things you can try next.
Create a ruleset and assign it your assistant in order to tailor its behavior and personality to your preferences.
Create a tool and assign it to your assistant in order to give it additional capabilities, such as dynamic web scraping.
Deploy a sample structure to perform tasks that go above and beyond basic chatting, such as converting PDFs to CSVs using natural language, comparing GPT against Claude and Gemini, or orchestrating multi-agent workflows.