LocalAI

LocalAI 4.9.0 introduces significant improvements in authentication, chat context compression, and UI organization. It also expands model support with MiniMax-H3 video generation and Qwen3-TTS on llama.cpp, alongside performance enhancements for downloads and backend operations.

This release enhances the model gallery with fallback mechanisms and introduces the NVIDIA NeMo-Speech.cpp backend. It also includes dependency updates and CI fixes.

This release includes several bug fixes for GGUF metadata, vLLM-cpp backend loading, and UI stability. It also adds new Qwen model variants to the gallery and enhances vLLM-cpp configuration.

LocalAI 4.8.0 introduces significant new modalities like vllm.cpp (alpha), 3D generation, and audio.cpp, alongside performance improvements, enhanced gallery features, and hardened distributed mode.

This release includes a bug fix related to the injection of llama.cpp serving options. The change ensures these options are only applied when the path is specifically for llama.cpp.

LocalAI 4.7.0 introduces significant enhancements in content generation, including UI-managed voice cloning, local video and avatar generation, and interleaved reasoning with tool calls. It also adds new audio engines and improves reliability across various features.

This release updates dependency versions, refactors internal code using slices.Contains, and adds support for MiniCPM models.

This patch release focuses on stability and resource management, introducing a new API endpoint for model capabilities and fixing several issues related to authentication, UI presentation, and VRAM utilization for context and embeddings.

LocalAI 4.6.0 is a reliability-focused release that brings full-speed GPU offload for AMD ROCm backends and predictable realtime session warmups. It also introduces conversation forking in the UI and hardens distributed worker lifecycle management.

This release focuses heavily on improving distributed system stability, enhancing model backend capabilities with new GGML detectors, and fixing numerous bugs related to configuration, imports, and streaming retries. Key features include cross-replica state synchronization and new voice/face detection backends.

This release focuses on fixing critical CI build breaks for several backends and improving compatibility for Whisper on Darwin systems. It also introduces one new model via the gallery agent.

This patch release fixes a specific issue related to how the RUN_BINARY is derived on Darwin systems.

This release focuses on improving packaging and stability, particularly for macOS builds and various backend components like piper, silero-vad, and llama-cpp. It also includes fixes for distributed configuration updates.

This minor release primarily addresses a build issue for the opus backend on macOS systems.

This release focuses heavily on expanding hardware compatibility, particularly adding Darwin/Metal (MPS) support across many backends, alongside significant UI enhancements and various bug fixes related to configuration and distributed processing.

LocalAI 4.5.0 introduces major new capabilities including depth perception, sound event classification, on-device TTS, and NER-based PII filtering. The release also enables continuous batching by default and ships a fully redesigned React UI.

This release introduces significant new TTS capabilities with OmniVoice and expanded Piper voices, alongside performance improvements in request routing and fixes for Darwin builds. It also adds Indonesian and Korean language support.

This patch updates underlying dependencies including llama.cpp and CrispASR, and resolves critical bugs related to vLLM tool argument parsing and CUDA environment setup.

This release introduces streaming capabilities for the realtime pipeline stages and fixes a compatibility issue with recent vLLM versions.

LocalAI 4.4.0 is a major multimodal and distributed release, introducing two new audio backends (Parakeet and CrispASR), native object detection, video understanding/generation, and a new intelligent middleware layer for routing and PII filtering.

This release introduces the NVIDIA NeMo Parakeet ASR backend and enhances security by hardening outbound HTTP client redirects. It also includes several dependency updates.

This release introduces the ability to honor per-request reasoning effort for chat completions and includes several bug fixes related to tool-call streaming and application shutdown handling.

This release focuses on a specific bug fix within the turboquant component related to guarding grpc-server fields for forks.

This release primarily updates several underlying dependencies, including llama.cpp, whisper.cpp, and stable-diffusion.cpp, and includes a fix for handling content in openresponses.

This release introduces significant new features like model routing, PII filtering, and object detection backends, alongside numerous dependency updates and bug fixes across distributed systems and model handling.