zhongkaifu / TensorSharp

What it does

• Local models, several interfaces. One engine for CLI use, browser chat and Ollama/OpenAI-compatible APIs, across managed CPU and native accelerator backends.
• Text and multimodal models. Dense and MoE GGUF models, reasoning, image/audio input and document questions. See Supported models for each family's capabilities.
• Embeddings and media generation. EmbeddingGemma 2 provides text/code embeddings and optional image, audio and video inputs. Generate and edit images with Qwen-Image-2.1 and Turbo, or generate video with stereo audio using LTX 2.5 Distilled
• Editable video projects. Video Studio in WebUX Chat and TensorAgent shares a persistent multitrack editor, native generation jobs, reviewed agent proposals, undo/redo and local FFmpeg export. Long videos are made as chains of takes, each st
• Efficient serving. Continuous batching and a paged, Radix prefix-shared KV cache are on by default. Speculative decoding and multi-GPU placement are available for supported models. See Features.