Grove: The Syntactic Layer for Efficient AI Agent Navigation
Stop wasting tokens on whole-file reads. Grove provides precise, byte-level symbol navigation for coding agents using a lightweight syntactic layer.
Lemonade 11.5.0: Mastering Intelligent Model Routing and AI Workflows
Discover Lemonade 11.5.0: New intelligent model routing policies, server-side job engines, ModelScope support, and MCP client capabilities for local AI.
Scaling Real-Time Transcription: A Guide to WhisperLive
Master real-time speech-to-text with WhisperLive. Learn how to deploy Faster Whisper, TensorRT, and OpenVino backends for high-performance AI streaming.
pi-task: Deterministic Spec-Orchestration for Local Models
Stop local LLM drift and hallucinations. Discover pi-task, the deterministic pipeline for reliable, crash-safe, and verifiable agentic workflows.
Meetily: The Privacy-First Alternative to Cloud-Based Meeting Assistants
Secure your meetings with Meetily. A 100% local AI assistant for real-time transcription and summaries using Whisper and Ollama. No cloud, no bots.
Vibe: The Privacy-First Desktop App for Offline Transcription
Transcribe audio and video locally with Vibe. A cross-platform desktop app for private transcription using Whisper and Ollama.
Beyond the Wall of Text: Mastering Transcription with WhisperX
Upgrade OpenAI Whisper with WhisperX for word-level timestamps, speaker diarization, and up to 70x faster transcription speeds.
whisper.cpp: Exciting High-Performance Offline Speech-to-Text
Run OpenAI's Whisper model fully offline. Learn how to use whisper.cpp for high-performance, hardware-accelerated transcription on Apple Silicon and NVIDIA.
Colibri: Running a 744B Parameter Model on 25GB of RAM
Discover Colibri: a tiny C engine that streams experts from disk to run massive 744B parameter models like GLM-5.2 on machines with only 25GB of RAM.
Running Frontier MoE Models on Consumer Hardware via vLLM-Moet
Deploy frontier MoE models like GLM-5.2 (753B) on consumer Blackwell GPUs using 2-bit experts and FP4 recovery with the vLLM-Moet patch.