DarkspyreStudios 

Darkspyre's source-compatible Microsoft.ML.Tokenizers fork with corrected SentencePiece optional-token handling.

Operating system credential vault for .NET. One interface over macOS Keychain, Windows Credential Manager, and the Linux Secret Service, with in-memory and encrypted file stores for hosts that have no vault.

Shared typed contracts for Darkspyre Connect and Tensor inference implementations.

Storage-neutral blob persistence for .NET, with an atomic filesystem provider rooted in the operating system's per-user application-data directory.

Shared model-pack, persistence, backend, and execution infrastructure for Darkspyre Tensor inference.

Universal conversation and embedding contracts, shared provider infrastructure and automatic Connect discovery.

The one GGML native runtime owner that every Tensor engine running on GGML borrows.

Qwen GGUF in-process LLM inference through TensorSharp.

Supra2-IMG in-process image generation through TensorSharp.

In-process text embeddings through TensorSharp GGUF sentence encoders.

Resident EmbeddingGemma text and image inference through ONNX Runtime CPU and WebGPU.

Connects hosts and MCP programs to the host-tools local service, which it bundles.

Universal conversations through the Antigravity CLI provider.

UI-neutral editor state and host contracts shared by the Darkspyre inference setup editors.

Registration, validation and ordered discovery of the Darkspyre Tensor engine implementations.

Host-placed Avalonia setup editor for Darkspyre inference providers.

Universal conversations through the Codex provider.

Universal conversations and embeddings through the OpenAiApi provider.

Resident CPU/CUDA Whisper speech transcription with host-owned model storage and lifetime.

Image, speech and audio inference connectors for Darkspyre applications.