Back to Details

ai-say

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

AI Say

AI Say

Local text-to-speech via 
~/.local/bin/ai-say
. Kokoro-first with flite and speech-dispatcher fallbacks. Requires the voice stack to be installed first.
本地文本转语音功能通过
~/.local/bin/ai-say
实现。优先使用Kokoro TTS，同时支持flite和speech-dispatcher作为回退方案。使用前需先安装语音栈。

Prerequisites

前置要求

The voice stack must be installed before using this skill:
bash
git clone https://github.com/Ema93sh/ai-voice-stack.git
cd ai-voice-stack
scripts/voice-up-setup.sh --with-system-deps
This installs 
ai-say
to 
~/.local/bin/
and sets up Kokoro TTS.
使用本技能前必须先安装语音栈：
bash
git clone https://github.com/Ema93sh/ai-voice-stack.git
cd ai-voice-stack
scripts/voice-up-setup.sh --with-system-deps
此操作会将
ai-say
安装到
~/.local/bin/
并配置Kokoro TTS。

Usage

使用方法

Speak text:
bash
~/.local/bin/ai-say "Hello world"
Pipe text:
bash
echo "Read this aloud" | ~/.local/bin/ai-say
Temporary voice override:
bash
AI_KOKORO_VOICE=am_michael ~/.local/bin/ai-say "Different voice"
朗读文本：
bash
~/.local/bin/ai-say "Hello world"
管道输入文本：
bash
echo "Read this aloud" | ~/.local/bin/ai-say
临时切换语音：
bash
AI_KOKORO_VOICE=am_michael ~/.local/bin/ai-say "Different voice"

Available Voices

可用语音

English (American)

美式英语

Female: 
af_alloy
, 
af_aoede
, 
af_bella
, 
af_heart
(default), 
af_jessica
, 
af_kore
, 
af_nicole
, 
af_nova
, 
af_river
, 
af_sarah
, 
af_sky
Male: 
am_adam
, 
am_echo
, 
am_eric
, 
am_fenrir
, 
am_liam
, 
am_michael
, 
am_onyx
, 
am_puck
, 
am_santa
女声： 
af_alloy
, 
af_aoede
, 
af_bella
, 
af_heart
（默认）, 
af_jessica
, 
af_kore
, 
af_nicole
, 
af_nova
, 
af_river
, 
af_sarah
, 
af_sky
男声： 
am_adam
, 
am_echo
, 
am_eric
, 
am_fenrir
, 
am_liam
, 
am_michael
, 
am_onyx
, 
am_puck
, 
am_santa

English (British)

英式英语

Female: 
bf_alice
, 
bf_emma
, 
bf_isabella
, 
bf_lily
Male: 
bm_daniel
, 
bm_fable
, 
bm_george
, 
bm_lewis
女声： 
bf_alice
, 
bf_emma
, 
bf_isabella
, 
bf_lily
男声： 
bm_daniel
, 
bm_fable
, 
bm_george
, 
bm_lewis

Other Languages

其他语言

Spanish (
ef_dora
, 
em_alex
), French (
ff_siwis
), Hindi (
hf_alpha
, 
hf_beta
, 
hm_omega
, 
hm_psi
), Italian (
if_sara
, 
im_nicola
), Japanese (
jf_alpha
, 
jf_gongitsune
, 
jf_nezumi
, 
jf_tebukuro
, 
jm_kumo
), Portuguese (
pf_dora
, 
pm_alex
), Chinese (
zf_xiaobei
, 
zf_xiaoni
, 
zf_xiaoxiao
, 
zf_xiaoyi
, 
zm_yunjian
, 
zm_yunxi
, 
zm_yunxia
, 
zm_yunyang
)
西班牙语（
ef_dora
, 
em_alex
）、法语（
ff_siwis
）、印地语（
hf_alpha
, 
hf_beta
, 
hm_omega
, 
hm_psi
）、意大利语（
if_sara
, 
im_nicola
）、 日语（
jf_alpha
, 
jf_gongitsune
, 
jf_nezumi
, 
jf_tebukuro
, 
jm_kumo
）、 葡萄牙语（
pf_dora
, 
pm_alex
）、中文（
zf_xiaobei
, 
zf_xiaoni
, 
zf_xiaoxiao
, 
zf_xiaoyi
, 
zm_yunjian
, 
zm_yunxi
, 
zm_yunxia
, 
zm_yunyang
）

Persistent Voice Change

永久切换语音

Edit 
~/.config/ai-audio.env
:
bash
export AI_KOKORO_VOICE="${AI_KOKORO_VOICE:-af_bella}"
编辑
~/.config/ai-audio.env
文件：
bash
export AI_KOKORO_VOICE="${AI_KOKORO_VOICE:-af_bella}"

Volume / Gain

音量 / 增益

If speech is too quiet:
bash
export AI_KOKORO_GAIN_DB=18
~/.local/bin/ai-say "Volume test"
如果语音音量太小：
bash
export AI_KOKORO_GAIN_DB=18
~/.local/bin/ai-say "Volume test"

Diagnostics

诊断工具

Check stack health:
bash
~/.local/bin/voice-status
Check audio devices:
bash
pactl list short sinks
pactl list short sources
cat ~/.config/ai-audio.env
Test direct tone on a specific sink:
bash
ffmpeg -hide_banner -loglevel error -f lavfi -i 'sine=frequency=880:duration=3' -f wav - | paplay --device='<sink-name>'
检查语音栈健康状态：
bash
~/.local/bin/voice-status
检查音频设备：
bash
pactl list short sinks
pactl list short sources
cat ~/.config/ai-audio.env
在指定输出设备上测试直接音调：
bash
ffmpeg -hide_banner -loglevel error -f lavfi -i 'sine=frequency=880:duration=3' -f wav - | paplay --device='<sink-name>'

Subcommands

子命令

Install

安装

Run the full voice stack installer. Finds the repo automatically or clones it:
bash
bash scripts/install.sh --with-system-deps
运行完整的语音栈安装程序。会自动查找仓库或进行克隆：
bash
bash scripts/install.sh --with-system-deps

Doctor

诊断

Check installation health — reports PASS/FAIL for every component:
bash
bash scripts/doctor.sh
检查安装健康状态——会报告每个组件的通过/失败状态：
bash
bash scripts/doctor.sh

Notes

注意事项

  • ai-say
    is Kokoro-first, with flite and spd-say fallback paths.
  • Text is truncated at 1600 chars and chunked at 320 chars for reliable playback.
  • Keep messages respectful and follow user intent exactly for spoken content.
  • ai-say
    优先使用Kokoro TTS，同时支持flite和spd-say作为回退方案。
  • 文本会被截断为1600字符，并按320字符分段以保证可靠播放。
  • 朗读内容需保持得体，并严格遵循用户的需求。