English Text to Speech
Free AI English text to speech online. 28 American and British English voices, local text processing, WAV and MP3 export, and no signup.
Try it now — no signup or per-character charge
Browser audio synthesis; text handling and network needs depend on the selected engine and language.
Our English text-to-speech tool runs entirely in your browser using the Kokoro TTS model. With 20 American English voices and 8 British English voices, you can find the perfect tone for any project.
Key features:
- 28 English voices (20 US, 8 UK) with quality ratings from A to D
- Top picks: Heart (A, warm storytelling), Bella (A-, energetic vlogs), Emma (B-, professional British), Michael (B, clear reviews)
- English text processing and audio synthesis stay in the browser after required assets load
- The first visit and uncached model or voice files still require network access
- Export as WAV for studio-quality audio or MP3 for compressed delivery
- Up to 50,000 characters per session with automatic chunking for long texts
- WebGPU acceleration is available on supported Chrome and Edge configurations
Best for: YouTube voiceovers, podcast intros, audio books, e-learning materials, and accessibility.
American English voices include Heart (A), Bella (A-), Nicole (A), Sarah (A-), Nova (B+), Kore (B+), Jessica (B+), Sky (B), River (B), and more. These voices are optimized for North American pronunciation and natural prosody.
British English voices include Emma (B-), Alice (B+), Isabella (B+), Lily (B+), Daniel (B+), George (B-), Lewis (B-), Fable (B-). These voices deliver authentic British pronunciation with proper rhythm and intonation for UK-focused content.
Tips for best English TTS results:
- Use proper punctuation — commas add pauses, periods create stops, question marks raise pitch
- Compare WebGPU and WASM on your own device because generation speed varies by hardware and browser
- For long texts, break into paragraphs with clear punctuation for natural rhythm
- The q4 model (~305MB) offers the best quality-to-size balance for most users
- Multiple voices can be combined for dialogue or interview-style narration
Whether you need a warm voice for business content or a conversational tone for a creative project, test the exact script and keep the selected engine, voice, and speed with the export.
Sponsored
Ads help keep OfflineTTS free to use.
American and British English Are Separate Voice Sets
The English page combines twenty American and eight British Kokoro presets, but a flag or friendly name does not guarantee one regional pronunciation for every word. Test proper nouns, dates, acronyms, currencies, and domain-specific terms in the exact voice you plan to use. If a script mixes US and UK spelling, decide which reading convention the audience expects before generating a long file.
English Kokoro uses local phonemization after the model and selected voice assets are available, so the entered script is not sent to the non-English OfflineTTS phonemization endpoint. The first load, model host, website hosting, and ordinary analytics still involve network requests. Clear browser site data to remove cached assets, and delete WAV or MP3 exports separately from the device.
Review English Names, Numbers, and Chunk Boundaries
A useful evaluation passage includes a person and place name, an abbreviation, a decimal, a year, a currency amount, a quotation, and one long sentence. Compare several voices with the same q4 or fp32 model, speed, browser, and backend. Catalog grades and traits are internal discovery labels rather than a controlled benchmark, so a lower-sorted preset may fit a particular script better.
For long narration, keep paragraphs coherent and listen across each automatic chunk join. Check for repeated words, clipped consonants, inconsistent volume, or pauses that do not match the text. WAV is the safer intermediate for editing; MP3 is smaller but lossy. Keep the reviewed source script and voice ID with the export so a correction can be reproduced.
Why Use Our English Text to Speech
Local Text Processing
English Kokoro phonemization and audio synthesis run in the browser after required model and voice assets load.
Works Offline
After model and voice assets are cached, English generation can work without a network connection until those assets are cleared.
No Per-Character Charge
No API key or signup is required; the current tool accepts up to 50,000 characters in one generation workflow.
WAV & MP3 Export
Download as WAV for studio-quality audio or MP3 for compressed output. Compatible with all editors.
Popular Use Cases
🎬 YouTube Voice-Overs
Add professional English narration to videos without recording equipment. Top voices: Heart (warm), Bella (energetic), Michael (professional).
🎙️ Podcast Production
Generate intros, outros, and full episodes with consistent voice quality. Mix multiple voices for interview-style segments.
📚 Audiobook Narration
Convert manuscripts to audiobooks with natural-sounding voices. Export as WAV for post-production.
🎓 E-Learning & Accessibility
Add voice narration to courses and make content accessible to visually impaired users across English-speaking audiences.
Available English Voices
| Voice | Type | Best For | Preview |
|---|---|---|---|
 Heart | A | Warm, natural — best for storytelling and educational content | |
 Bella | A- | Expressive and dynamic — great for vlogs and creative projects | |
 Emma | B- | Professional British English — ideal for business and formal narration | |
 Michael | B | Clear, professional male voice — perfect for reviews and tutorials |
How It Works
Paste Text
Enter your english text (up to 50,000 chars)
Choose Voice
Pick from english voices
Generate
AI creates speech on your device
Download
Save as WAV or MP3
English Text to Speech — FAQ
Yes. OfflineTTS does not charge per character and requires no signup or API key. Audio synthesis runs in your browser with the engine you select.
Offline behavior depends on the selected engine and language. Supertonic, Piper, Kitten, and English Kokoro can synthesize locally after their model files download. Non-English Kokoro uses the OfflineTTS phonemization service before audio is generated on your device.
Audio synthesis runs locally in your browser. Fully local-capable engine and language combinations keep the input on your device after model download. Non-English Kokoro sends plain text to the OfflineTTS phonemization service and receives pronunciation data before local synthesis.
Start Generating English Speech Now
No signup, no per-character fee, and browser-based audio synthesis.
Open TTS Tool →