Structured outputs on Amazon Bedrock: Schema-compliant AI responses
Machine Learning Blog
This article announces structured outputs on Amazon Bedrock, enabling schema-compliant JSON responses from foundation models through constrained decoding.
- Eliminates JSON parsing failures, missing fields, type mismatches, and schema violations
- Two mechanisms: JSON Schema output format for data extraction; strict tool use for agentic workflows
- Constrained sampling ensures responses always match specified JSON schema structure
- Grammar compilation cached for 24 hours, improving performance for repeated schemas
- Supports JSON Schema Draft 2020-12 subset with enums, arrays, objects, and string formats
- Requires additionalProperties set to false; recursive schemas and external references unsupported
- Available across all commercial AWS Regions for nine model providers including Anthropic, OpenAI, Google
- Use cases span financial services, healthcare, ecommerce, legal, and customer service industries
Structured outputs eliminates custom validation logic, enabling reliable production-ready AI applications with guaranteed schema compliance.
The AWS News Feed is currently looking for gold sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.
Related articles
Feb 4
2026
2026
Structured outputs now available in Amazon Bedrock
Jun 26
2025
2025
Structured data response with Amazon Bedrock: Prompt Engineering and Tool Use
Jan 15
2026
2026
Build a generative AI-powered business reporting solution with Amazon Bedrock
Nov 7
2025
2025
Introducing structured output for Custom Model Import in Amazon Bedrock
The AWS News Feed is currently looking for silver sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.