Frontier CoreML audio models in your apps — text-to-speech, speech-to-text, voice activity detection, and speaker diarization. In Swift, powered by SOTA open source.
FluidAudio is a Swift SDK that runs local audio AI models on Apple devices for speech recognition, text-to-speech, voice activity detection, and identifying who is speaking. It is intended for developers building iOS and macOS apps with low-latency audio processing. The catalogue add-ons provide workflows for using the SDK.
Latest release v0.18.2 — v0.18.2 — CAM++ v2, Pocket TTS 2026-10, long-form ASR repairs · 1 Nov 2026
7 files for Claude Code, GitHub Copilot, Codex and OpenCode: apple-neural-performance-expert, code-search-ast-grep, Stop, FluidAudio and 3 more — 3,845 tokens loaded in every session.
.github/copilot-instructions.md A 552 tok AGENTS.md A 551 tok CLAUDE.md A 2,742 tok .claude/settings.json A — .claude/settings.json A — .claude/agents/apple-neural-performance-expert.md A — .claude/agents/code-search-ast-grep.md A — These files are FluidInference/FluidAudio's own configuration — they tell Claude Code, GitHub Copilot, Codex and OpenCode how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.