Release 15.0.8

• feat(dsp): Separate structured output example input fields with newlines and allow missing required fields during structured output validation in response processing. 6150f36

Release 15.0.7

• feat: enhance structured output handling with distinct extraction modes and improved prompt rendering for complex fields 7ad07fe

Release 15.0.6

• feat: Enhance complex object and JSON extraction, add validation tests, and improve error messages with LLM output. 100ed60

Release 15.0.5

• feat: add documentation for AWS Bedrock, Vercel AI SDK, and Ax Tools packages. 95962ae
• feat(anthropic): add validation for arbitrary json objects in structured outputs (#459) 7db81c5

Release 15.0.4

• fix: build issue 71b5ae8
• fix(anthropic): add anthropic-beta header for web-search on Vertex AI (#457) df13f8c
• fix(ace): Ensure only input fields are passed to curator (#456) 8c0c13f
• feat(mipro): Expand MIPROv2 optimizer to tune instructions and examples (#453) 2f3e6ac
• fix: allow f.object().array() as input field (#452) d36ddd6
• Add support for gemini-embedding-001 model (#449) 8372b13

Release 15.0.3

• feat: Update Anthropic schema cleaning to preserve default, oneOf, anyOf, allOf and conditionally remove additionalProperties. dbc419c
• docs: Clarify assertion error handling in examples, distinguishing between AxAssertionError for retries and Error for immediate failure. 6396ae7

Release 15.0.2

• feat: Implement and document parallel function calling for Google Gemini. cb1a310
• ci(deps): bump actions/setup-node from 4 to 6 (#405) 983af6b
• chore(deps): bump @opentelemetry/exporter-metrics-otlp-http (#416) 60b2c6a
• chore(deps): bump ai from 5.0.56 to 5.0.81 (#417) 1dfb29b

Release 15.0.1

• feat: Introduce AxSignature.hasComplexFields() for consistent complex type detection and update example documentation. b1dc107

Release 15.0.0

• feat: Introduce structured (XML) prompt generation with format protection and tests, and remove individual streaming result logging. f04c787

Release 14.0.44

• feat(anthropic): update and align Vertex AI model maxTokens values (#426) f042d7b