ChindaTTS Text-to-Speech

Speaks Thai or mixed Thai and English text in the Kaitom voice or a voice cloned from a short clip, streamed back as 24 kHz PCM. These are the V3 endpoints; the V1 and V2 endpoints are deprecated.

Alpha

The service is in alpha and may be intermittently unavailable.

POST/v3/store/audio/tts

+1 endpoints

Try it

• Price1 ICper 400 characters
• Served2.1Mcalls

• v3.0
• Active

Input

text

สวัสดีครับ น้องไข่ต้ม เวอร์ชั่น 3

speed

1

OutputPOST /v3/store/audio/tts

HTTP/1.1 200 OK
Content-Type: application/octet-stream

<raw PCM, signed 16-bit little-endian, mono, 24 kHz, streamed>
duration_seconds = byte_length / 48000

Every field, under ResponseOther endpoints: /clone see Request

Request

curl -X POST https://api.iapp.co.th/v3/store/audio/tts \
  -H "apikey: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text": "สวัสดีครับ น้องไข่ต้ม เวอร์ชั่น 3", "speed": 1.0}' \
  --output speech.pcm

Response

200OK. The fields are explained under Response.

HTTP/1.1 200 OK
Content-Type: application/octet-stream

<raw PCM, signed 16-bit little-endian, mono, 24 kHz, streamed>
duration_seconds = byte_length / 48000

Request

JSON for the default voice; multipart/form-data for cloning.

Headers

apikeystringrequired

Your API key, created under API Keys in the control panel.

Body

textstringrequired

The text to speak, up to 10,000 characters. About 1,000 Thai characters per call works best; longer text is split on the server and takes longer.

speedfloat

Default voice only: the rate, default 1.0; 0.8 to 1.2 sounds natural.

filefile

Cloning only, required there: 10 to 20 seconds of clean, mono Thai speech, WAV or MP3. A shorter or noisy clip is refused with 400.

consentboolean

Cloning only, required there: true to confirm the voice is yours or its speaker has agreed. Without it the request is refused with 400.

Endpoint	Body	Does	Price
/v3/store/audio/tts	text, speed	Speaks in the Kaitom voice	1 IC per 400 characters
/v3/store/audio/tts/clone	text, file, consent	Speaks Thai in the cloned voice; the clip is used for that request only and not stored	1 IC per 400 characters

The older paths /v3/store/audio/tts/tts and /v3/store/audio/tts/tts/clone reach the same two speech endpoints.

Response

/v3/store/audio/tts streams raw PCM, signed 16-bit little-endian, mono, 24 kHz, as application/octet-stream; duration_seconds = byte_length / 48000. To play or save it, add a WAV header or run ffmpeg -f s16le -ar 24000 -ac 1 -i speech.pcm speech.wav. /v3/store/audio/tts/clone returns a WAV file (audio/wav), ready to play.

/v3/store/audio/tts streams raw PCM, signed 16-bit little-endian, mono, 24 kHz, as application/octet-stream; duration_seconds = byte_length / 48000. To play or save it, add a WAV header or run ffmpeg -f s16le -ar 24000 -ac 1 -i speech.pcm speech.wav. /v3/store/audio/tts/clone returns a WAV file (audio/wav), ready to play.

Status codes

Code	Meaning
400	Malformed JSON
413	Text over 10,000 characters
503	Temporarily unavailable; retry later

Details

Limits

• Thai and English for the default voice; Thai only for cloning.
• 100 requests per second per API key.
• Punctuation sets the pauses. To fix how a brand or acronym is said, write it in Thai script: ไอแอป, not iApp.

Data handling

Text and reference clips are processed in memory and not kept after the response. The service is GDPR and PDPA compliant.

Deprecated endpoints

The V1 and V2 endpoints still answer, at the same price, but are no longer maintained. Move to /v3/store/audio/tts.

Endpoint	Was
/v3/store/speech/text-to-speech/kaitom, also /thai-tts-kaitom2/tts	V2, Kaitom voice
/v3/store/speech/text-to-speech/kaitom/v1, also /thai-tts-kaitom/tts	V1, Kaitom voice
/v3/store/speech/text-to-speech/cee, also /thai-tts-cee/tts	V1, Cee voice

To move, send JSON with text and speed in place of form fields or a query string, and drop language: the language is detected. The answer is raw PCM, to be wrapped as WAV before it plays (see Response), not an audio file. V3 has no Cee voice.

Changelog

Version	Changes
ChindaTTS	The V3 endpoints run on ChindaTTS, with the same request and response. The V1 and V2 endpoints are deprecated; see Deprecated endpoints.
V3	/v3/store/audio/tts replaces /v3/store/speech/text-to-speech/kaitom, with a JSON body, the language detected instead of a language mode, up to 10,000 characters, 24 kHz output, text normalisation for numbers, dates, currency, times and percentages, and voice cloning.