Select Language:
If you’re considering using Amazon Bedrock for your project, understanding how the pricing works can help you plan your budget better. Here’s a simple guide to breaking down the costs based on what many users need—setting up Knowledge Bases with PDFs stored in an S3 bucket.
First, there’s the cost associated with initial indexing or ingestion. When you upload your PDFs, you’ll pay depending on the parser you choose. If you opt for Bedrock Data Automation (BDA), you’ll be charged per page. The rate is $0.01 per page, so if you’re processing 1,000 pages, your cost for indexing will be around $10. Keep in mind, if you choose Foundational Model parsers instead, the charges are based on the number of tokens used for input and output, rather than pages.
Next, there are query costs when you actually start using the Knowledge Base. Every question or request you make consumes tokens—these include the tokens from your input (your question) and from the system’s response. The price varies depending on which Foundation Model you select for querying.
For ongoing use, think about how many queries you’ll handle regularly, whether you’ll need a fixed, predictable performance level, and how frequently your knowledge base needs updating. These factors will influence your choices and costs.
When planning your budget, be sure to evaluate a few points:
- The size of your knowledge base
- The expected frequency of queries
- The initial costs, including setting up your knowledge base
- The ongoing operational expenses
- The performance needs, which might mean choosing between on-demand or provisioned throughput
With Amazon Bedrock, you’re only charged for what you actually use, so there are no hidden costs. Your AWS account is set up to automatically include all services, but you’ll only pay for the resources you utilize.
If you want to get detailed pricing info, visit the official AWS pages for Amazon Bedrock. They provide up-to-date details on costs and help you figure out the best option for your project.




