English | Japanese

yomi - Nostr/Bluesky/Misskey.io/Discord TTS Reading Client

A Nostr/Bluesky/Misskey.io/Discord client that reads out posts using text-to-speech.

For Users

Usage

1. Enable and configure your platforms:
   • Nostr: Use NIP-07 extension (nos2x for Chrome, nos2x-fox for Firefox) or enter pubkey manually
   • Bluesky: Settings > Privacy and Security > App Passwords > Enter handle and app password
   • Misskey.io: Settings > Service integration > Generate access token
   • Discord: Create a Discord Bot (see discord-bot/README.md) > Launch the bot with your token and channel ID > Enter WebSocket URL in yomi settings

     cd discord-bot
     npm install
     node bot.cjs --token <BOT_TOKEN> --channel <CHANNEL_ID>

2. Open yomi in your browser
3. Click the gear icon to open the config dialog
4. Click "Start" to begin reading posts from your follows

Requirements

Windows/Mac/iOS/Android

• A modern browser (Chrome, Firefox, Safari, Edge)

Linux

• A modern browser with Web Speech API support
• Speech synthesis voices installed:

Ubuntu/Debian:

sudo apt install speech-dispatcher speech-dispatcher-espeak-ng espeak-ng

Arch:

sudo pacman -S speech-dispatcher espeak-ng

Tested Environments

OS	Browser	Foreground Playback	Background Playback	Bluetooth Playback	Speech Recognition (on device)	Speech Recognition (Bluetooth)
Windows 11 25H2	Chrome 143.0.7499.194	✔	✔	✔	✔	✔
iOS 26.2	Safari 26.6	✔		✔	✔
Android 13	Chrome 143.0.7499.194	✔		✔	✔	✔
Android 13	Firefox Nightly 149.0a1	✔	✔	✔

Supported Languages

Available languages for text-to-speech depend on your browser and OS. Below are commonly supported languages:

Code	Language	Windows	Mac	Android	iOS
en	English	✓	✓	✓	✓
ja	日本語	✓	✓	✓	✓
zh	中文	✓	✓	✓	✓
ko	한국어	✓	✓	✓	✓
es	Español	✓	✓	✓	✓
fr	Français	✓	✓	✓	✓
de	Deutsch	✓	✓	✓	✓
it	Italiano	✓	✓	✓	✓
pt	Português	✓	✓	✓	✓
ru	Русский	✓	✓	✓	✓
ar	العربية	✓	✓	✓	✓
hi	हिन्दी	✓	✓	✓	✓
th	ไทย	✓	✓	✓	✓
vi	Tiếng Việt	✓		✓
nl	Nederlands	✓	✓	✓	✓
pl	Polski	✓	✓	✓	✓
tr	Türkçe	✓	✓	✓	✓
uk	Українська	✓		✓

Note: On Linux, available languages depend on what espeak-ng supports. You can install additional language packs to extend support.

Auto Language Detection

The "Auto detect by author" and "Auto detect by note" features use franc-min for language detection. franc-min supports the following 82 languages:

Code	Language	Code	Language	Code	Language
amh	አማርኛ	arb	العربية	azj	Azərbaycan
bel	Беларуская	ben	বাংলা	bho	भोजपुरी
bos	Bosanski	bul	Български	ceb	Cebuano
ces	Čeština	ckb	کوردی	cmn	中文
deu	Deutsch	ell	Ελληνικά	eng	English
fra	Français	fuv	Fulfulde	guj	ગુજરાતી
hau	Hausa	hin	हिन्दी	hms	苗语
hnj	Hmong	hrv	Hrvatski	hun	Magyar
ibo	Igbo	ilo	Ilokano	ind	Indonesia
ita	Italiano	jav	Jawa	jpn	日本語
kan	ಕನ್ನಡ	kaz	Қазақ	kin	Kinyarwanda
koi	Коми	kor	한국어	lin	Lingála
mad	Madura	mag	मगही	mai	मैथिली
mal	മലയാളം	mar	मराठी	mya	မြန်မာ
nld	Nederlands	npi	नेपाली	nya	Chichewa
pan	ਪੰਜਾਬੀ	pbu	پښتو	pes	فارسی
plt	Malagasy	pol	Polski	por	Português
qug	Kichwa	ron	Română	run	Kirundi
rus	Русский	sin	සිංහල	skr	سرائیکی
som	Soomaali	spa	Español	srp	Српски
sun	Sunda	swe	Svenska	swh	Kiswahili
tam	தமிழ்	tel	తెలుగు	tgl	Tagalog
tha	ไทย	tur	Türkçe	ukr	Українська
urd	اردو	uzn	Oʻzbek	vie	Tiếng Việt
yor	Yorùbá	zlm	Melayu	zul	isiZulu
zyb	壮语

Note: For auto language detection to work correctly, the detected language must also be supported by TTS.

For Developers

Requirements

• Node.js 18+
• npm

Build

npm install
npm run build

Run

npm run dev

Open http://localhost:5173 in your browser.

Architecture

Flow from fetching posts to speech synthesis:

flowchart TB
    subgraph Sources["Post Fetching"]
        Nostr["Nostr<br/>subscribeToNotes()"]
        Bluesky["Bluesky<br/>getTimeline()"]
        Misskey["Misskey<br/>subscribeToTimeline()"]
        Discord["Discord<br/>connectStream()"]
        Test["Test<br/>window.testpost()"]
    end

    subgraph Queue["Queue Management"]
        NoteWithRead["NoteWithRead<br/>{id, pubkey, content,<br/>created_at, read, source}"]
        NotesRef["notesRef / setNotes"]
    end

    subgraph Processing["Reading Process (processNextNote)"]
        FindUnread["Find oldest unread note"]
        MarkRead["Mark as read: true"]
        FormatAuthor["Format author name<br/>(remove emoji, etc.)"]
        ProcessText["processTextForSpeech()<br/>(handle emoji/URL/Nostr address)"]
        BuildText["fullText =<br/>authorName + content"]
    end

    subgraph Speech["Speech Synthesis"]
        SpeechManager["speechManager.speak()"]
        WebSpeechAPI["Web Speech API"]
    end

    Nostr --> NoteWithRead
    Bluesky --> NoteWithRead
    Misskey --> NoteWithRead
    Discord --> NoteWithRead
    Test --> NoteWithRead
    NoteWithRead --> NotesRef
    NotesRef --> FindUnread
    FindUnread --> MarkRead
    MarkRead --> FormatAuthor
    FormatAuthor --> ProcessText
    ProcessText --> BuildText
    BuildText --> SpeechManager
    SpeechManager --> WebSpeechAPI
    WebSpeechAPI -->|onEnd| FindUnread

Loading

WebSocket Debug Mode

Forward all console output to a WebSocket server for remote debugging.

1. Start the WebSocket server:

node tools/wsserver.cjs > console-log.txt

2. Open the app with ?ws=host:port parameter:

http://localhost:5173/?ws=localhost:8080

License

• This project: MIT License
• Third-party libraries: THIRD_PARTY_LICENSES.md