Introduction
Send the text to be read and, optionally, one of nine voices, a short note on delivery and a language. The task returns one MP3 file (24 kHz mono) with the script read in that voice. instructions takes plain-language direction such as tone, emotion or pace and is not read aloud. SpicyAPI applies no content filter: NSFW scripts are read as written. You are billed per 1,000 characters of the text you send, with a minimum of 100 characters.
| Field | Value |
|---|---|
| Input | Text |
| Output | Audio |
| Required field | text — 1 to 2,000 characters |
| Optional fields |
|
| Result | One MP3 file, 24 kHz, mono |
| Billed quantity | Characters of text sent, per 1,000; minimum 100 |
| Typical time | About 30 seconds for 300 characters |
| Retention | Generated audio is kept for 14 days |
Request parameters
| Parameter | Required | Description |
|---|---|---|
text | Yes | Write Qwen3 TTS’s spoken script in text, including the punctuation that should create pauses in the reading. at least 1 characters · at most 2000 characters |
voice_id | No | One of nine preset voices. Defaults to Serena. |
instructions | No | Plain-language direction on tone, emotion or pace. Shapes the delivery and is not spoken. at most 1000 characters |
language | No | Language of the reading. auto detects it from the script. |
Pricing
$0.068 / 1,000 characters
Qwen3 TTS calculates the charge from each 1,000 characters of script before synthesis, so the pre-run quote is the final charge for that text.
Use cases
Character and companion voices
Nine distinct voices — from bright and edgy to low and mellow — each with a short note on how they talk, kept the same for a whole series.
Chinese, Japanese and Korean lines
Dylan and Eric speak Beijing and Sichuan dialects, and Ono_Anna and Sohee are native Japanese and Korean speakers, so lines in those languages sound natural rather than accented.
Uncensored audio stories
NSFW scripts are read as written. In our tests the model spoke every word of the monologues and narration we sent without skipping or cutting the clip short.
Voice tracks for video
Generate the line here, then hand the MP3 to a talking-avatar or lip-sync model so a face performs it.
Prompt guide & pro tips
Send the script in text
Write Qwen3 TTS’s spoken script in text, including the punctuation that should create pauses in the reading.
Pick a voice
voice_iddefaults to Serena, a warm, gentle female voice. Ryan and Aiden are English male voices; Ono_Anna and Sohee are Japanese and Korean; Dylan and Eric speak Beijing and Sichuan dialects.Describe the delivery
Put a sentence or two in instructions — for example 'breathless and teasing, close to the microphone'. It is not read aloud.
What it returns
One MP3 file per task, 24 kHz mono. Download and archive the output of script reading with Qwen3 TTS’s nine preset voices within 14 days; SpicyAPI deletes its stored copy after that period. The audio returned by Qwen3 TTS contains the performance, with no accompanying word-level timings or subtitle file.
Choosing a voice
voice_id takes one of nine preset voices and defaults to Serena:
Serena— warm, gentle young femaleVivian— bright, slightly edgy young femaleUncle_Fu— seasoned male with a low, mellow timbreRyan— dynamic male with strong rhythmic driveAiden— sunny American maleOno_Anna— playful Japanese femaleSohee— warm, emotional Korean femaleDylan— youthful male speaking Beijing dialectEric— lively male speaking Sichuan (Chengdu) dialect
Directing the delivery
instructions takes a sentence or two of plain language about how to read: tone, emotion, pace. In Qwen3 TTS, delivery instructions direct the performance while text supplies the spoken words; the direction itself is not read aloud. Speed and pitch sliders are not exposed; ask for them in instructions.
Language
language defaults to auto, which detects the language from the script. Set zh, en, de, fr, es, it, pt, ru, ja or ko when a short or mixed script is read in the wrong one.
Billing
Billed per 1,000 characters of text, counting spaces, punctuation and line breaks, with a minimum of 100 characters per request; instructions is free. Qwen3 TTS calculates the charge from each 1,000 characters of script before synthesis, so the pre-run quote is the final charge for that text. Failed requests are not charged.
Content
SpicyAPI adds no content filter. SpicyAPI adds no prompt or result filter to script reading with Qwen3 TTS’s nine preset voices; this describes the platform’s treatment of Qwen3 TTS, with your responsibility for lawful use unchanged.
SpicyAPI adds no prompt or result filter to script reading with Qwen3 TTS’s nine preset voices; this describes the platform’s treatment of Qwen3 TTS, with your responsibility for lawful use unchanged. Secure the rights to scripts and source recordings used for script reading with Qwen3 TTS’s nine preset voices, and identify generated speech as synthetic rather than attributing it to a real speaker.
API quick start
# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
-H "Authorization: Bearer $SPICY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
-d '{
"model": "alibaba/qwen-3-tts/text-to-speech",
"input": {
"text": "Welcome back. Your files finished rendering a minute ago and are ready to download.",
"language": "auto",
"voice_id": "Serena"
}
}'
# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
-H "Authorization: Bearer $SPICY_API_KEY"
See the API reference for authentication, the complete schema, polling, and error handling.





