v0.91.0

What's Changed

• feat(docs): Boost Guidelines by @mysticseagull in #652
• Passthrough aspect ratio to gemini flash image request map by @patrickbrouwers in #651
• docs: update guidelines link by @ijpatricio in #649
• chore: Create .gitattributes by @ijpatricio in #648
• docs(openai): fix formatting by @cosmastech in #624
• fix(doc): Correct the namespace of Prism facades in samples by @chuoke in #620
• fix(xai): stream handling with 0 value mid stream by @patrickbrouwers in #636
• fix(gemini): stream handler sends empty array for generationConfig by @hipig in #630
• fix(openrouter): Handle FinishReason::Length by @michabbb in #633
• bug(gemini) Set default values for groundingSupport indices by @aimeos in #642
• feat(openai) parallel tool calls provider option by @patrickbrouwers in #650
• feat(core) Extend Media class methods to allow passing mime type by @aimeos in #643

New Contributors

• @mysticseagull made their first contribution in #652
• @ijpatricio made their first contribution in #649
• @chuoke made their first contribution in #620
• @michabbb made their first contribution in #633

Full Changelog: v0.90.0...v0.91.0

v0.90.0

What's Changed

• fix: ensure tool result output is always string by @Shoghy in #599
• feat(openai): Multi-image edit support by @sixlive in #640

Breaking Changes

• Image edits for OpenAI are no longer passed via withProviderOptions and are now passed via Image value objects in the withPrompt's additionalContent parameter.
• OpenAI mask option now utilizes the Image value object instead of a raw resource

New Contributors

• @Shoghy made their first contribution in #599

Full Changelog: v0.89.0...v0.90.0

v0.89.0

What's Changed

• feat(openai): support file citations by @rasmuscnielsen in #608
• Prevent recursive endless loop on create a resource from url media by @patrickbrouwers in #609
• feat(mcp): Adds mcp_servers client option support for Anthropic by @Arkitecht in #610

New Contributors

• @rasmuscnielsen made their first contribution in #608
• @patrickbrouwers made their first contribution in #609
• @Arkitecht made their first contribution in #610

Full Changelog: v0.88.1...v0.89.0

v0.88.1

What's Changed

• docs: Add ObjectSchema requirement warnings for structured output by @richardhowes in #597
• Fix typo in exception message for max tokens by @freekmurze in #606

New Contributors

• @richardhowes made their first contribution in #597
• @freekmurze made their first contribution in #606

Full Changelog: v0.88.0...v0.88.1

v0.88.0

What's Changed

• fix(gemini): streaming output tool yields by @sixlive in #596
• feat(gemini): yield metadata & usage information by @avengerweb in #449
• feat(openrouter): add support for provider routing by @mboisvertdupras in #578

New Contributors

• @avengerweb made their first contribution in #449
• @mboisvertdupras made their first contribution in #578

Full Changelog: v0.87.1...v0.88.0

v0.87.1

What's Changed

• bug(gemini) Fix schema generation for structured content by @aimeos in #589
• fix(openai) Duplicate reasoning with tool calls by @sixlive in #595

Full Changelog: v0.87.0...v0.87.1

v0.87.0

What's Changed

• docs(favicon): add favicon to website by @grpaiva in #579
• feat(citations)!: citations abstraction + openai web search annotations by @ChrisB-TL in #483
• Update dev dependencies & fix failing tests by @sixlive in #591
• feat(Gemini): Support rate limit and overload exception by @aimeos in #585

Breaking Changes

• Anthropic citations now saved on additionalContent under the 'citations' key and use the new MessagePartWithCitations and Citation value objects
• Gemini search groundings are now saved on additional content under the 'citations' key and use the new MessagePartWithCitations and Citation value objects

New Contributors

• @grpaiva made their first contribution in #579

Full Changelog: v0.86.0...v0.87.0

v0.86.0

What's Changed

• feat(openrouter): Improve OpenRouter reasoning and structured response handling by @robtesch in #561
• docs: update OpenAI provider documentation by @christeyerl in #571
• feat(gemini): add image generation support by @beefan in #560
• feat(openai): gpt-5 structured support by @sixlive in #573
• fix(openai): rate limits by @sixlive in #574
• fix(tools): error handling by @sixlive in #576

New Contributors

• @robtesch made their first contribution in #561
• @christeyerl made their first contribution in #571
• @beefan made their first contribution in #560

Full Changelog: v0.85.0...v0.86.0

v0.85.0

What's Changed

• feat(ollama): add support for thinking chunks in streaming responses (gpt-oss) by @skatecram in #548
• refactor: Remove response messages by @sixlive in #516
• feat(ollama): support think parameter by @sixlive in #558
• feat(elevenlabs): Speech to Text by @sixlive in #557

Breaking Changes

• Removes responseMessages from the response objects

Full Changelog: v0.84.0...v0.85.0

v0.84.0

What's Changed

• refactor(openai): GPT-5 adjustments by @sixlive in #551

Full Changelog: v0.83.2...v0.84.0