Text to Speech – Voice Output in Scratch

🔊 Text to Speech – Voice Output in Scratch #

The Text to Speech extension brings realistic spoken voice to your Scratch projects.
Make your characters talk, narrate stories, create voice assistants, build language learning tools, and add accessibility features – all with simple blocks that convert text into natural-sounding speech.
Works in 23 languages with 5 unique voices. ✨

---

🌟 Overview #

• Natural Speech Synthesis: Convert any text into spoken words with realistic voices.
• 23 Languages Supported: Speak in English, Spanish, French, German, Chinese, Arabic, and many more.
• 5 Unique Voices: Alto, Tenor, Squeak, Giant, and Kitten – each with distinct characteristics.
• Simple Integration: Just one block to make your sprites talk!
• Language Detection: Automatically defaults to your editor’s language.
• Cloud-Based: Uses Scratch’s synthesis service for high-quality speech generation.

---

✨ Key Features #

• 5 distinct voice personalities: neutral (Alto/Tenor), funny (Squeak/Giant), and playful (Kitten).
• 23 supported languages with natural pronunciation.
• Automatic language matching based on editor locale.
• Gender-appropriate voice selection per language.
• Character-based voices for creative storytelling.
• Sync with Scratch animations and sound effects.
• Perfect for accessibility, language learning, and interactive narratives.

---

🚀 How to Use #

1. Go to: pishi.ai/play
2. Open the Extensions section.
3. Select the Text to Speech extension.
4. Use the “speak [WORDS]” block to make your sprite talk!
5. (Optional) Set voice and language before speaking.
6. Your sprite will speak the text out loud – Scratch waits until speech finishes.

Tips

• The extension automatically detects your editor’s language and uses it by default.
• Use different voices for different characters to create conversations.
• Combine with “think” or “say” blocks to display text while speaking.
• Works best with internet connection – requires cloud speech synthesis service.
• Keep text under 128 characters for best performance.

---

🧱 Blocks and Functions #

 

🗣️ Main Speech Block #

speak [WORDS]

Converts text to speech and plays it as audio.
The script waits until the speech finishes before continuing – perfect for synchronized storytelling.

How it works:

• Type or insert text to be spoken.
• The block sends the text to Scratch’s synthesis service.
• Audio is generated with the selected voice and language.
• Audio plays through your speakers/headphones.
• The block waits until speech completes before moving to the next block.

Arguments:

• [WORDS] – Text to speak (string, up to 128 characters)

Examples:

• speak [Hello, my name is Scratch Cat!]
• speak [join [The answer is ] (score)] – Speak dynamic text with variables
• speak [What is your favorite color?]

Note: This block waits for speech to complete – use it for dialogue, narration, and synchronized animations.

 

---

🎭 Voice Selection #

set voice to [VOICE]

Changes the voice character for speech synthesis.
Each voice has a unique personality – perfect for different characters or moods.

Available Voices:

• alto – Neutral, ambiguous gender voice (default) – professional, clear
• tenor – Neutral, ambiguous gender voice – slightly deeper than alto
• squeak – High-pitched, playful voice – great for small characters, excited speech
• giant – Low-pitched, deep voice – perfect for large characters, serious tones
• kitten – Ultra-high baby voice – speaks “meow” for all words (fun character effect!)

Examples:

• set voice to [alto] → Standard neutral voice
• set voice to [squeak] → High-pitched, energetic voice
• set voice to [giant] → Deep, serious voice
• set voice to [kitten] → Cat character (says “meow meow meow”)

Creative Uses:

• Use alto for narrators, teachers, or professional characters.
• Use tenor for heroes, leaders, or confident characters.
• Use squeak for fairies, children, robots, or excited emotions.
• Use giant for monsters, villains, authority figures, or serious moments.
• Use kitten for pet characters or comedic effect.

 

---

🌍 Language Selection #

set language to [LANGUAGE]

Changes the language for speech synthesis.
The extension automatically defaults to your editor language, but you can override it for multilingual projects.

Supported Languages (23 total):

• English (en) – American English pronunciation
• Spanish (European) (es) – Castilian Spanish
• Spanish (Latin American) (es-419) – Latin American Spanish
• French (fr) – French pronunciation
• German (de) – German pronunciation
• Italian (it) – Italian pronunciation
• Portuguese (Brazilian) (pt-br) – Brazilian Portuguese
• Portuguese (European) (pt) – European Portuguese
• Chinese (Mandarin) (zh-cn) – Mandarin Chinese (simplified & traditional)
• Japanese (ja) – Japanese pronunciation
• Korean (ko) – Korean pronunciation
• Arabic (ar) – Modern Standard Arabic
• Hindi (hi) – Hindi pronunciation
• Russian (ru) – Russian pronunciation
• Dutch (nl) – Dutch pronunciation
• Polish (pl) – Polish pronunciation
• Turkish (tr) – Turkish pronunciation
• Danish (da) – Danish pronunciation
• Swedish (sv) – Swedish pronunciation
• Norwegian (nb) – Norwegian Bokmål
• Icelandic (is) – Icelandic pronunciation
• Romanian (ro) – Romanian pronunciation
• Welsh (cy) – Welsh pronunciation

How it works:

• The extension automatically sets language to match your Scratch editor’s language (if supported).
• Use this block to override the default or create multilingual projects.
• Language setting is saved with the project.
• Some languages have only female voices available – the extension automatically adjusts voice pitch.

Examples:

• set language to [Spanish (European)] → Speak in Spanish
• set language to [French] → Speak in French
• set language to [Japanese] → Speak in Japanese

Language Learning Projects:

• Create projects that teach pronunciation in different languages.
• Build multilingual vocabulary quizzes.
• Make story translations with spoken audio.
• Practice language listening comprehension.

Note on Single-Gender Languages:
Some languages (Arabic, Chinese, Hindi, Korean, Norwegian, Romanian, Swedish, Turkish, Welsh) only have a female voice available from the synthesis service. The extension automatically adjusts pitch for Tenor and Giant voices in these languages.

 

---

🎓 Educational Uses #

• Accessibility: Add voice narration for visually impaired users or non-readers.
• Create read-aloud story projects – combine text display with speech.
• Build language learning tools – teach pronunciation, vocabulary, phrases.
• Make interactive tutorials with spoken instructions.
• Develop voice-responsive games and quizzes.
• Create multilingual projects for diverse classrooms.
• Teach character voice and dialogue writing.
• Build assistive technology projects for students with reading challenges.

---

🎮 Example Projects #

• Talking Storybook: Characters narrate story text with different voices.
• Language Tutor: Speak vocabulary words and phrases in different languages – combine with Speech Recognition extension.
• Voice Assistant: Scratch cat acts as a helpful AI assistant answering questions.
• Multilingual Greeter: Greet users in multiple languages based on selection.
• Quiz Game with Spoken Questions: Voice reads questions aloud for accessibility.
• Interactive Dialogue: Create conversations between multiple sprites with different voices.
• Audio Notifications: Use voice to announce scores, achievements, timer updates.
• Pronunciation Practice: Display word, speak it, ask user to repeat.
• Character Conversations: Use Alto for one character, Tenor for another, Squeak for a third.
• Comedy Sketch: Use Giant for the villain, Squeak for the hero, Kitten for the sidekick.

🧩 Try it yourself: pishi.ai/play

 

---

🔧 Tips & Troubleshooting #

 

🔊 Text to Speech Specific Tips #

• No sound? Check your device volume and browser sound permissions. Make sure speakers/headphones are connected.
• Speech not working? This extension requires an internet connection – the synthesis service is cloud-based.
• Wrong language pronunciation? Make sure you’ve set the correct language before speaking. Language defaults to editor locale.
• Text too long? Keep text under 128 characters – longer text is automatically truncated.
• Voice not changing? Make sure to use “set voice to [VOICE]” before the “speak” block.
• Kitten voice speaking gibberish? That’s normal! Kitten voice replaces all words with “meow” for character effect.
• Speech sounds weird in some languages? Some languages only have female voices – Tenor/Giant may sound unusual due to pitch adjustment.
• Want to stop speech early? Use the red stop button or “stop all” block – all speech stops immediately.
• Multiple sprites talking? Set different voices for each sprite to create conversations.
• Sync speech with animation? Use “speak” in sequence with costume changes, movement, or sound effects.
• Speech overlapping? “speak” block waits for completion – use multiple “speak” blocks in sequence for dialogue.

---

💡 Creative Tips #

🎭 Voice Characterization #

• Narrator: Use Alto or Tenor for clear, professional narration.
• Children/Small Characters: Use Squeak for high-pitched, energetic speech.
• Villains/Large Characters: Use Giant for deep, commanding presence.
• Animals/Pets: Use Kitten for playful cat characters (all words become “meow”).
• Robots/AI: Use Tenor or Alto with technical vocabulary.
• Fantasy Creatures: Experiment with Squeak or Giant for unique character voices.

📖 Storytelling Techniques #

• Combine “say” blocks with “speak” blocks to show and speak text simultaneously.
• Use different voices for different story characters.
• Add pauses between speech with “wait” blocks for dramatic effect.
• Alternate between narrator voice (Alto/Tenor) and character voices (Squeak/Giant).
• Use variables to store dialogue and speak dynamic responses.

🌍 Multilingual Projects #

• Create language switcher with buttons or variables.
• Store translations in lists – speak the appropriate translation.
• Build “Learn a Language” projects with word pronunciation.
• Make international greeting projects – say “hello” in many languages.

---

🔒 Privacy and Safety #

• Text sent to speech synthesis is processed by Scratch’s cloud service (synthesis-service.scratch.mit.edu).
• No text is permanently stored – it’s only used to generate audio.
• Generated audio is streamed directly to your browser – not saved on servers.
• Language and voice settings are saved in your project file locally.
• Requires internet connection to function – speech synthesis is not offline.

---

🧪 Technical Info #

• Synthesis Service: Scratch’s cloud-based text-to-speech service
• Server URL: synthesis-service.scratch.mit.edu
• Timeout: 10 seconds (requests exceeding timeout will fail)
• Max Text Length: 128 characters (automatically truncated)
• Audio Volume: 250% (boosted for clarity)
• Speech Format: Audio buffer (decoded and played through Scratch audio engine)
• Languages Supported: 23 languages with natural pronunciation
• Voices: 5 distinct voices with playback rate adjustments
• Single-Gender Languages: Arabic, Chinese, Hindi, Korean, Norwegian, Romanian, Swedish, Turkish, Welsh
• Internet Required: Yes – cloud synthesis service

---

🆚 Voice Comparison Chart #

↔ Swipe left or right to view full table on mobile

---

🔗 Related Extensions #

• 🎤 Speech Recognition – convert voice to text (perfect complement for voice interactions)
• 🌐 Translate – translate text between languages before speaking
• 💬 ChatGPT – generate dialogue with AI, then speak it with Text to Speech
• 🎵 Music – combine speech with music and sound effects

---

📚 Learn More #

• AI Extensions Help at pishi.ai
• Scratch Wiki: Text to Speech Extension
• Scratch Project Ideas

---