All chunks fire simultaneously · Direct + Proxy fallback · Zero server ffmpeg

Convert text to
natural speech.

All chunks fired simultaneously — your device does the merging. Unlimited text length, lightning fast, no queue.

580+

Voices

∞

Text length

∞×

True parallel

Server ffmpeg

Your Text

Voice Effects

Pitch 0

DeeperNormalHigher

Speed 0

SlowerNormalFaster

Generate

🎤 No voice selected — pick one →

Preparing...

Ready to play

Download MP3

Follow our WhatsApp Channels

Voice Library

Loading voices...

API Documentation REST · v1

Base URL

How it works

The API exposes two lightweight endpoints. GET /api/voices returns the voice list. POST /api/tts synthesizes a single chunk (max 1950 chars) and returns raw audio/mpeg bytes. For long text: split client-side → fire all chunks in parallel with Promise.all() → concatenate the Blobs. MP3 is a sequential stream so Blob concat works perfectly — zero server-side ffmpeg needed.

Endpoints

GET/api/voices

Returns all available voices with 1-based index numbers, grouped by language. Use the index in TTS calls.

Response: { "success": true, "total": 580+, "voices": [ { "index": 1, "id": "voice-107", "name": "Andrew Multilingual", "gender": "Male", "language": "Multilingual", "country": "United States" }, ... ], "grouped": { "Multilingual": [...], "English": [...] } }

POST/api/tts

Synthesizes one chunk of text and returns raw audio/mpeg binary. Max 1950 chars per call. No timeout set — Vercel's 30s function limit applies.

voiceIndexnumberrequired— 1-based index from /api/voices

textstringrequired— Text to synthesize (max 1950 chars per call)

pitchnumberoptional— Pitch adjustment from -100 (deeper) to 100 (higher), default 0

ratenumberoptional— Speaking speed from -100 (slower) to 100 (faster), default 0

POST /api/tts Content-Type: application/json { "voiceIndex": 1, "text": "Hello, this is a test.", "pitch": 10, "rate": -5 } ← 200 Content-Type: audio/mpeg (raw binary) Headers: X-Pitch, X-Rate, X-Voice-Name, X-Char-Count

Parallel processing (how this UI works)

// 1. Split text at sentence boundaries into ≤1950-char chunks const chunks = splitText(text, 1950); // 2. Fire ALL chunks simultaneously — no sequential waiting const blobs = await Promise.all( chunks.map(chunk => fetch('/api/tts', { method: 'POST', headers: { 'Content-Type': 'application/json' }, body: JSON.stringify({ voiceIndex, text: chunk, pitch, rate }) }).then(r => r.blob()) ) ); // 3. Concat Blobs — MP3 is a stream, this just works const merged = new Blob(blobs, { type: 'audio/mpeg' }); const url = URL.createObjectURL(merged);

Error codes

400

Missing or invalid parameter (invalid pitch/rate range, etc.)

405

Method not allowed — use GET / POST

502

TTS provider error — retry

SpeechSter APIREST · v1 · Free

Base URL

GET/api/voicesjson

List all TTS voices with index, name, language, gender.

Click Open → see live response in your browser

Open

POST/api/ttsjson

Convert text to natural speech. Returns audio stream.

Param	Type		Description
text	string	required	Text to speak.
voiceIndex	number	required	Voice index from /api/voices.
pitch	number	optional	Pitch (default 0).
rate	number	optional	Speed (default 0).

Copy & run in your terminal

curl -X POST {ORIGIN}/api/tts \
  -H "Content-Type: application/json" \
  -d '{"text":"Hello from Ahm7xMakki","voiceIndex":0}' \
  --output speech.mp3

POST endpoints can't be opened by clicking a URL — paste the curl above into your terminal, or call it with fetch() in code.

Convert text tonatural speech.

Convert text to
natural speech.