Show HN: TalkiTo – enabling voice and Slack for Claude Code and Codex CLI (github.com)

TalkiTo is an open-source toolkit that brings voice, Slack, WhatsApp and SMS I/O to code-focused LLMs — currently Claude Code and OpenAI Codex — via a CLI, browser extension and Python library. Developers can install via pip or the repo, run commands like talkito claude/codex, or launch a web SSE server for extension use. It also supports remote triggers (Slack channels, Twilio SMS/WhatsApp with zrok tunneling) so long-running commands and outputs can be delivered to messaging channels. Technically, TalkiTo is designed to be provider-agnostic: pluggable TTS (OpenAI voices, AWS Polly, Azure, Google, ElevenLabs, KittenTTS, Kokoro) and ASR (Whisper local via faster-whisper/pywhispercpp with int8/int8_float16/coreml options, Google, Azure, Deepgram, AssemblyAI, AWS Transcribe, Houndify). It supports on-device lightweight TTS (KittenTTS, Kokoro) for privacy/offline use, streaming transcription providers for low-latency dictation, and configurable env files (.env/.talkito.env). Requires Python 3.10+, platform audio libs (portaudio/say/espeak) and exposes independent tts/asr modules for reuse. Licensed AGPLv3, TalkiTo lowers friction for hands‑free coding, accessibility, and remote automation workflows by connecting code CLIs and code-aware LLMs to natural voice and messaging interfaces.