Ollama adds local speech and decision models.
Local LLM runner adding macOS speech synthesis and vLLM semantic router d3 decision models.
Brings audio streaming and multi-size classification models directly to local desktop inference.
TypeScript AI toolkit fixing non-streaming Responses API structured output failures.
Enables OpenAI and Azure Responses models to select final answers while preserving conversation history.
Local inference engines are expanding beyond text generation into speech synthesis and specialized decision routing, changing how teams integrate multimodal capabilities directly into local hardware architectures; teams should test it in a reversible pilot before adoption.
The same card posted on social. Open it for the full size.
