Summary
read_text is the entire surface, which makes this a speaker bolted onto an assistant that otherwise only writes.
Its parameters are the entire configuration — text, plus speed, tone, volume and voice — and because they ride on each call rather than sitting in a config file, a narration loop can change register without restarting anything. Anything about voice selection beyond those parameters belongs to Bouyomi-chan rather than to this server.
What it is
A single-tool bridge to Bouyomi-chan: read_text hands over a line of text to be spoken, and carries the delivery settings with it.
What you get
- read_text speaks the text you pass through Bouyomi-chan.
- Delivery is set per call rather than globally: speed, tone, volume and voice are parameters on the same call, so one line can be read back differently from the next.
Requirements
Bouyomi-chan itself, since read_text speaks through it. No credentials are configured.
Setup effort
One command — npx -y github:uraoz/bouyomichan-mcp-nodejs