TeamAI's Datastores feature allows users to create embeddings of documents, webpages, and websites, enabling powerful search and chat capabilities. This article is specifically tailored for our legacy Pro and Sumo users who need to use their own resources to set up and utilize Datastores.
Prerequisites
To use Datastores, you'll need:
An OpenAI API key with access to the embedding model
A Qdrant vector database cluster (hosted on Qdrant or self-hosted)
Step 1: Obtain Your OpenAI API Key
Log in to https://platform.openai.com/
Choose the "API" option from the main menu
Navigate to "API Keys" in the sidebar
Copy your OpenAI API key
Add the API key to the advanced settings of your TeamAI workspace
Step 2: Set Up Your Qdrant Vector Database Cluster
Sign up at https://cloud.qdrant.io/ if you haven't already
After logging in, select an existing Qdrant cluster or create a new one from the right sidebar
Fill in the required details for your new cluster
You'll be redirected to the list of your clusters, where you'll see your new cluster in the "Creating" state
Once the cluster is ready and in a "Healthy" state, click on "API Key" under the cluster name
Create a new API key and link it to your new cluster
Copy the API key (note that it's only copyable at the time of creation)
Return to your cluster and copy the endpoint URL from the "Endpoint" box
Paste both the API key and endpoint URL into the advanced settings of your TeamAI workspace
With your OpenAI API key and Qdrant vector database cluster set up, you're now ready to use Datastores. Start uploading files, webpages, and websites, and enjoy the powerful search and chat capabilities provided by TeamAI.
Troubleshooting Tips
Ensure your OpenAI API key has access to the embedding model
Double-check that you've correctly copied and pasted the API keys and endpoint URL
If you encounter issues with your Qdrant cluster, reach out to Qdrant support for assistance
FAQs
Q: Can I use other vector databases besides Qdrant?
A: Currently, TeamAI only supports Qdrant for storing embeddings. However, we're always exploring integrations with other vector databases to provide more options for our users.
Q: What if I don't have an OpenAI API key?
A: To use Datastores, you'll need an OpenAI API key with access to the embedding model. If you don't have one, sign up for an OpenAI account and create a new API key with the necessary permissions.