Deploy Agent with Knowledge Base
Introduction
Congratulations! You’ve built your agent with a Knowledge Base and published it to Studio. Now it’s time to deploy it to Sema4.ai Control Room and make it available to your team.
Deploying agents with a Knowledge Base follows the standard agent deployment process — with a few extra considerations specific to KB setup.
Step-by-step guide
Initiate deployment
In Sema4.ai Control Room, navigate to the Gallery section and find your agent. Click on the Deploy button to start the deployment process.
Review details
Once you initiate the deployment process, you will be asked to review the details of the agent. You’ll see all your queries from the action package.
The view will show the agent’s data sources — including the pgvector storage and semantic Knowledge Base defined in your action package.
Each agent can currently include one Knowledge Base.
Configure Knowledge Base
You will be prompted to configure the Knowledge Base settings. Either select an existing Knowledge Base or create a new one. If creating a new one, specify:
- Name: A unique identifier for your Knowledge Base
- Description: A brief overview of what the Knowledge Base contains
- Vector Storage: Choose the vector database (e.g., pgvector) where embeddings will be stored
- Model: Select the embedding model to use for content ingestion
Deploy and verify
After configuration, deploy the agent. Once deployed, verify that the Knowledge Base is accessible and that queries return expected results.
✅ Your agent is now live with its Knowledge Base!