What's Changed
- Handle empty LLM response in do_run/1 (thinking model streaming fix) by @lurielle-studio in #484
- fix: skip Azure keepalive events in streaming responses by @vasspilka in #485
- feat: Langchain.Trajectory for easier evaluation of agents by @vasspilka in #481
- fix: add additionalProperties: false to tool schemas by @brainlid in #490
- Support multimodal tool results in ChatVertexAI by @brainlid in #491
- fix: reduce excessive Logger.error usage across the library by @brainlid in #492
- test: add regression tests for malformed tool call JSON loop (#443) by @brainlid in #493
- feat: add logprobs and top_logprobs support to ChatOpenAI by @brainlid in #494
- fix: handle Mistral responses without tool_calls key by @brainlid in #495
- prep for v0.6.3 by @brainlid in #496
New Contributors
- @lurielle-studio made their first contribution in #484
Full Changelog: v0.6.2...v0.6.3