Skip to main content
POST
/
v1
/
tts
/
slng
/
inworld
/
max:1.5
curl --request POST \
  --url https://api.slng.ai/v1/tts/slng/inworld/max:1.5 \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "text": "Hello, this is a test of Inworld text to speech.",
  "voice": "Ashley",
  "modelId": "inworld-tts-1.5-max"
}
'
{
  "audioContent": "SUQzBAAAAAAAI1RTU0UAAAAPAAADTGF2ZjYwLjE2LjEwMAAA",
  "usage": {
    "processedCharactersCount": 48,
    "modelId": "inworld-tts-1.5-max"
  }
}

Authorizations

​
Authorization
string
header
required

API key issued by SLNG. Pass as Authorization: Bearer <token>.

Headers

​
X-Region-Override
enum<string>

Target region override. Auto-selected if not provided.

Available options:
us-east-1

Body

application/json

Inworld Max 1.5 synthesis request, SLNG-hosted. Audio output is configured with flat fields (encoding, sample_rate, bit_rate, speaking_rate) that the gateway maps to Inworld's audioConfig.

​
text
string
required

Text to synthesize. Max 2,000 characters.

Required string length: 1 - 2000
​
voice
string
default:Ashley

Inworld voice ID. Voices are multilingual — each has a native language but can speak any supported language via the language parameter (best results when language matches the voice's native language). See the full catalog at https://docs.slng.ai/voices/inworld.

​
modelId
enum<string>
default:inworld-tts-1.5-max

ID of the Inworld TTS model.

Available options:
inworld-tts-1.5-max
​
language
enum<string>

BCP-47 language tag specifying the language the voice should speak the text in. Optional — when omitted, Inworld uses the voice's original prompt and auto-detects the language from the text. Voices are multilingual; best results when this matches the voice's native language. These are the 15 core languages supported by inworld-tts-1.5-max (inworld-tts-2 additionally supports ~85 experimental languages). An invalid code returns an error.

Available options:
ar-SA,
de-DE,
en-US,
es-ES,
fr-FR,
he-IL,
hi-IN,
it-IT,
ja-JP,
ko-KR,
nl-NL,
pl-PL,
pt-BR,
ru-RU,
zh-CN
​
deliveryMode
enum<string>

Only applies to inworld-tts-2. Controls output variation; ignored on Max 1.5.

Available options:
DELIVERY_MODE_UNSPECIFIED,
STABLE,
BALANCED,
CREATIVE
​
temperature
number
default:1

Higher values produce more expressive output; lower values more deterministic. Range (0, 2].

Required range: 0 < x <= 2
​
timestampType
enum<string>
default:TIMESTAMP_TYPE_UNSPECIFIED

Controls timestamp metadata returned with the audio. Adds latency.

Available options:
TIMESTAMP_TYPE_UNSPECIFIED,
WORD,
CHARACTER
​
applyTextNormalization
enum<string>
default:APPLY_TEXT_NORMALIZATION_UNSPECIFIED

Expands numbers, dates, and abbreviations before synthesis. Disabling may reduce latency.

Available options:
APPLY_TEXT_NORMALIZATION_UNSPECIFIED,
ON,
OFF
​
encoding
enum<string>
default:MP3

Output audio format. Maps to Inworld audioConfig.audioEncoding.

Available options:
LINEAR16,
MP3,
OGG_OPUS,
ALAW,
MULAW,
FLAC,
PCM,
WAV
​
sample_rate
enum<integer>
default:48000

Output sample rate in Hz. Maps to Inworld audioConfig.sampleRateHertz.

Available options:
8000,
16000,
22050,
24000,
32000,
44100,
48000
​
bit_rate
integer

Bits per second. Only applies to compressed formats (MP3, OGG_OPUS). Maps to Inworld audioConfig.bitRate.

​
speaking_rate
number
default:1

Playback speed. Values below 0.8 not recommended for quality. Maps to Inworld audioConfig.speakingRate.

Required range: 0.5 <= x <= 1.5

Response

Synthesis successful. Unlike other SLNG TTS models (which return raw audio bytes), Inworld Max responds with JSON: the request is SLNG-normalized but the response is Inworld's native body passed through unchanged, carrying base64-encoded audio plus usage and optional timestamp metadata.

Inworld Max 1.5 synthesis result. Inworld's native response, passed through by the gateway.

​
audioContent
string<byte>

Base64-encoded audio in the requested encoding. Max 16MB.

​
usage
object

Synthesis usage details.

​
timestampInfo
object

Timestamp metadata. Present only when timestampType is WORD or CHARACTER.