speak
GitHub调用chatterbox-nano模型将文本转换为语音，支持低延迟、高保真及声音克隆。适用于语音助手、游戏旁白、有声书等场景，按分钟计费且成本极低。
Trigger Scenarios
Install
npx skills add zerogpu/zerogpu-router --skill speak -g -y
SKILL.md
Frontmatter
{
"name": "speak",
"description": "Generate speech from text with chatterbox-nano, Resemble AI's 110M-parameter text-to-speech model, built for fast, high-quality speech generation at low latency, with expressive speech and voice cloning. Use for voice agents, conversational AI, gaming, accessibility, audiobooks, and content creation — anywhere the user asks to read text aloud, narrate it, or turn it into audio.",
"allowed-tools": "Bash(zerogpu chat_completions *)",
"argument-hint": "<text>"
}
Call chatterbox-nano. $ARGUMENTS is the raw text to speak — pass it verbatim, no escaping or quoting required (the heredoc below handles every shell metacharacter, newline, quote, and paren safely):
zerogpu chat_completions -m chatterbox-nano <<'ZGPU_END_OF_INPUT'
$ARGUMENTS
ZGPU_END_OF_INPUT
Output is the model's response as the API returns it. Report what was synthesized and relay the response's own reference to the audio — do not print an encoded audio payload back to the user.
At $0.01 per minute of generated audio it is the cheapest audio model on the platform and bills nothing per token. zerogpu chat_completions writes the response to stdout and takes no output-file option, so this skill does not save a playable file; if the user needs one, say so rather than reporting a path that does not exist. For the other direction, /zerogpu-router:transcribe is the speech-to-text route.
Savings note: only if the command output literally contains a line starting with 💰 ZeroGPU savings, append that exact line, unchanged, as the last line of your reply. If no such line is present, say nothing about savings and do not mention or suggest /zerogpu-router:cost-savings — this note is intentionally occasional, not shown every time.
Version History
- 605eceb Current 2026-11-02 19:13
