ArticleJakub Mroz, Henry NdubuakuAnnouncement
On-device speech recognition has historically required substantial memory and compute overhead. Whistle demonstrates that accurate multilingual transcription and voice-driven tool invocation can run locally on standard CPUs without cloud dependency or heavy runtimes.
By sharing structural layers with Needle, Cactus Compute minimises execution redundancy in the C++ engine. In reported benchmarks against Whisper base and Moonshine tiny v2, Whistle demonstrates notable latency and throughput advantages, alongside lower word error rates on benchmarks like LibriSpeech and FLEURS. The availability of prebuilt binaries across 17 architectures—ranging from iOS to RISC-V—makes it a versatile candidate for privacy-conscious edge applications.
Summary generated by AI (Gemini 3.8 Flash, 9 October 2026). It may contain errors – the original source is authoritative.
Similar in content, found by the AI search.
X post:Mario Zechner
Mario Zechner highlights Whistle by Cactus, an open speech-to-text model requiring only 16.9 MB of storage. It runs CPU-only without external dependencies and supports transcription in seven languages.
592bookmarks24.3kviews
Repository:google-ai-edge/LiteRT-LM
LiteRT-LM is an open-source inference orchestration layer from Google designed to run large language models directly on edge devices. It leverages GPU and NPU acceleration and powers on-device generative AI features in Chrome, Pixel Watch, and Chromebook Plus.
6.4kstarsC++
Repository:huggingface/speech-to-speech
Hugging Face has released speech-to-speech, a modular open-source framework designed to build voice agents with open-source models. The pipeline spans VAD, STT, LLM, and TTS, exposing the core OpenAI Realtime GA event set over WebSocket and WebRTC.
13kstarsPython
Article:Liquid AI
Liquid AI has launched LFM2.5-2.6B, an open-weight 2.6-billion parameter model engineered for local agentic workflows on phones, CPUs, and enterprise GPUs.
Repository:jamiepine/voicebox
Voicebox is a local-first, open-source voice studio combining text-to-speech, voice cloning, and system-wide dictation. It serves as a self-hosted alternative to proprietary cloud solutions such as ElevenLabs and WisprFlow.
55.6kstarsTypeScript
Repository:drumih/turbo-fieldfare
TurboFieldfare is a dedicated Swift and Metal runtime designed to run Google's Gemma 4 26B-A4B on Apple Silicon Macs within an approximately 2 GB RAM budget. It avoids loading the full 14.3 GB model into memory by keeping only the 1.35 GB core and KV cache in RAM while streaming the required mixture-of-experts weights from SSD on demand.
6.6kstarsSwift
Parts of this content were created with the assistance of AI.
