NobodyWho brings local LLMs to every device

// 3h agoINFRASTRUCTURE

NobodyWho brings local LLMs to every device

ANNOUNCEMENT PRODUCT GITHUB PRODUCT HUNT

NobodyWho is an open-source inference engine for running GGUF language, vision, speech-to-text, and text-to-speech models entirely on-device. It offers bindings for Swift, Kotlin, Flutter, React Native, Python, and Godot, with GPU acceleration and type-safe tool calling.

// ANALYSIS

NobodyWho makes local AI practical for developers building privacy-sensitive, offline-first apps across mobile, desktop, and game platforms. Its broad bindings are compelling, though performance still depends heavily on model size and device hardware.

–Supports thousands of GGUF models through llama.cpp and Hugging Face downloads
–Covers LLM inference, multimodal input, tool calling, STT, and TTS in one stack
–Metal and Vulkan acceleration extend local inference beyond desktop GPUs
–Cross-platform bindings reduce the need to maintain separate native inference integrations
–EUPL 1.2 permits commercial use but requires modified distributions to remain open source

// TAGS

nobodywhoinferenceedge-aiopen-sourcelocal-firstmultimodalspeech

DISCOVERED

3h ago

2026-08-20

PUBLISHED

9h ago

2026-08-20

RELEVANCE

9/ 10

AUTHOR

Pierre

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

TUTORIAL48m ago

Unlazy makes AI work prove completion

Unlazy is an open-source skill for Claude Code, Codex, Cursor, and other agents that decomposes complex tasks into work units with file-backed acceptance gates, runnable checks, and evidence. The video demonstrates how to adapt its workflow for parallel subagents.

OPEN SOURCE57m ago

Bright Data Skill Brings Scraping To Agent Terminals

This open-source skill teaches Claude Code, Cursor, Codex, and OpenCode to build, run, and self-heal Bright Data Scraper Studio scrapers from the terminal. It supports SERP research, competitor analysis, structured extraction, and reusable Collector IDs without maintaining scraper code.

OPEN SOURCE1h ago

Lazy Frames turns agents into video producers

Lazy Frames is an open-source, local-first toolkit that gives AI agents a spec-driven workflow for capturing assets, authoring scenes, adding audio, validating projects, previewing timelines, and rendering MP4s. Its deterministic pipeline produces byte-identical output from the same project and machine.