elbruno

.NET library that converts file formats to Markdown for AI pipelines, documentation workflows, and developer tools. Inspired by Python markitdown.

Structured document model for ElBruno.MarkItDotNet ingestion pipelines. Defines Document, Section, Block hierarchy for layout-aware processing.

Excel (.xlsx) to Markdown converter for ElBruno.MarkItDotNet. Converts spreadsheet sheets to Markdown tables.

PowerPoint (.pptx) to Markdown converter for ElBruno.MarkItDotNet. Converts slides and speaker notes to Markdown.

AI-powered converters for ElBruno.MarkItDotNet using IChatClient. Provides OCR, image captioning, and audio transcription.

A .NET library to download files (ONNX models, tokenizers, voice presets, etc.) from Hugging Face Hub repositories with progress reporting, caching, and authentication support.

Local embedding generation using Microsoft.Extensions.AI and ONNX Runtime

Run Qwen3-TTS text-to-speech locally from C# using ONNX Runtime. Models download automatically from HuggingFace on first run.

A .NET library for local text-to-image generation using Stable Diffusion and ONNX Runtime. Supports multiple models (SD 1.5, LCM Dreamshaper) with automatic model download from HuggingFace and native C# inference — no Python dependency.

Local LLM chat completions using Microsoft.Extensions.AI and ONNX Runtime GenAI. IChatClient implementation for running LLMs locally.

AI-powered knowledge graph builder for codebases

CPU runtime for ElBruno.Text2Image. Includes all text-to-image generation functionality with CPU-based inference via ONNX Runtime. Install this for environments without a GPU.

Microsoft Foundry cloud image generation for ElBruno.Text2Image. Provides FLUX.2 and MAI-Image-2 text-to-image generation via the Microsoft Foundry REST API. No local models needed — runs in the cloud.

NVIDIA CUDA GPU acceleration for ElBruno.Text2Image. Install this package instead of ElBruno.Text2Image to enable CUDA-based inference for Stable Diffusion image generation on NVIDIA GPUs.

Core domain types and storage interfaces for MemPalace.NET — local-first AI memory with wings/rooms/drawers hierarchy.

DirectML GPU acceleration for ElBruno.Text2Image. Install this package instead of ElBruno.Text2Image to enable DirectML-based inference for Stable Diffusion image generation on AMD, Intel, and NVIDIA GPUs (Windows only).

Microsoft.Extensions.VectorData integration for ElBruno.LocalEmbeddings

SQLite backend for MemPalace.NET with managed BLOB vector storage and cosine similarity search.

Windows systray monitor for Foundry Local — model load/unload notifications, mini status window

Temporal knowledge graph for MemPalace.NET with entity-relationship triples and validity windows.

Cross-platform CLI for AI text-to-image generation. Supports FLUX.2, MAI-Image-2 (Microsoft Foundry), GPT-Image-1.5, and GPT-Image-2 (Azure OpenAI). Lightweight cloud-first design (~30 MB). For local CPU/GPU inference, see ElBruno.Text2Image.Cli.Full.

Kernel Memory integration for ElBruno.LocalEmbeddings — provides ITextEmbeddingGenerator adapter and KernelMemoryBuilder extensions

Core abstractions for ElBruno.Speech — audio types, provider contracts, session interfaces, and error types.

RAG (Retrieval-Augmented Generation) pipeline for ElBruno.LocalLLMs. Provides document chunking, embedding storage, and semantic search.

CLIP-based image embedding generation using ONNX Runtime for local multimodal search and RAG scenarios

Audio primitives for ElBruno.Speech — WAV I/O, PCM conversion, resampling, framing, and ring buffers.

A .NET library for VibeVoice text-to-speech using ONNX Runtime. Supports automatic model download from HuggingFace, voice presets, and native C# inference with no Python dependency.

BitNet 1.58-bit LLM inference using bitnet.cpp. IChatClient implementation for Microsoft.Extensions.AI.

Voice cloning extension for ElBruno.QwenTTS. Clone any voice from a 3-second audio sample using the Qwen3-TTS Base model with ECAPA-TDNN speaker encoder.

A C# library for using Claude models (Sonnet 4.5, Haiku 4.5, Opus 4.1) deployed in Microsoft Foundry, implementing Microsoft.Extensions.AI abstractions

Local Whisper speech-to-text using ONNX Runtime. Auto-downloads models from HuggingFace and supports WAV plus raw PCM stream, byte, and float inputs.

Model downloader for ElBruno.LocalEmbeddings.ImageEmbeddings

Harrier-OSS-v1 embedding model integration for ElBruno.LocalEmbeddings — decoder-only transformer with last-token pooling and 94+ language support

Local audio transcription for ElBruno.MarkItDotNet using OpenAI Whisper via ONNX Runtime. Converts audio files to Markdown transcripts offline.

Intel Core Ultra NPU-accelerated local embedding generation using OpenVINO and ONNX Runtime

Speech pipeline orchestration for ElBruno.Speech — VAD → STT → LLM → TTS with bounded channels, barge-in, and session management.

Qualcomm Snapdragon X NPU-accelerated local embedding generation using QNN and ONNX Runtime

NPU-accelerated local embedding generation using DirectML and ONNX Runtime

Layout-aware and token-aware chunking for ElBruno.MarkItDotNet ingestion pipelines. Supports heading-based, paragraph-based, and token-aware strategies.

CLI tool for ElBruno.Speech — devices, transcribe, vad, talk, serve, and benchmark commands.

Windows-first .NET tool that forwards GitHub Copilot CLI events to the ElBruno.CopilotCLIMonitor systray app.

Windows system tray monitor for Aspire distributed applications. Install as a global tool and launch with `aspiremon`. See your Aspire resources at a glance, pin the ones you care about in a compact mini window with full URLs, and start/stop your AppHost from the tray.

ASP.NET Core integration for ElBruno.Speech — WebSocket endpoint, session registry, health checks, and protocol handling.

Command-line tool for converting files to Markdown. Powered by ElBruno.MarkItDotNet.

Silero VAD provider for ElBruno.Speech — ONNX Runtime inference, stateful streaming voice activity detection.

Windows audio I/O for ElBruno.Speech via NAudio — microphone input, speaker output, and device enumeration.

Windows-first Ollama monitor global tool launcher and configuration CLI.

OpenTelemetry instrumentation for ElBruno.Speech — activities, metrics, and Aspire-compatible setup.

Citation tracking and source traceability for ElBruno.MarkItDotNet ingestion pipelines. Provides page, block, and heading-path references for chunks.

Core abstractions for real-time audio conversation in .NET. Provides ITextToSpeechClient, IVoiceActivityDetector, and IRealtimeConversationClient interfaces following Microsoft.Extensions.AI patterns. Pluggable providers for STT, TTS, VAD, and LLM.