Skip to main content

Qwen-Audio 3.0 TTS Plus accepts a script and performance directions for expressive social voices, using 597 preset voices with configurable speed, pitch, volume and audio output.

From $0.019/1,000 characters

Workbench

Input

Text to speech
133/10000
3.0 TTS Plus script: required text, up to 10,000 characters, with no default script; its length determines the charge per 1,000 characters. at least 1 characters · at most 10000 characters

133 characters

3.0 TTS Plus casting: choose one of 597 preset voices belonging to this model; other-model IDs are rejected, no default is specified in the supplied facts, and selection does not affect price.
84/1000
3.0 TTS Plus direction: free-style emotion, tone, persona, pace or dialect notes that are not spoken or billed; no length limit or default direction is specified in the supplied facts. at most 1000 characters

Never read aloud. It only shapes how the script is performed.

Safety checkerOff

SpicyAPI does not run a safety checker on this model. Your prompt and media go to the model as written, and the model itself decides what comes back.

Checking your account before enabling submission and pricing.

Output

Ready for a new run
Example result
0:00 / –:––

Related models and next tasks

Missing the model you need?

Give every character a manner of speaking

Qwen-Audio 3.0 TTS Plus puts expression first for social and companion apps. Alibaba's 3.0 Plus offers 597 presets: 589 Chinese-language voices and eight English voices, with Luna Wang selected by default. Write the reply, describe the character's mood in instructions and generate a take at $0.019 per 1,000 text characters. SpicyAPI supplies no content filter, and our November 5, 2026 checks found that this endpoint spoke NSFW test scripts without rewriting them.

Open the playground
Give every character a manner of speaking
Preset cast
597 voices
Voice mix
589 Chinese-language + 8 English
Default speaker
Luna Wang, English
Reply capacity
10,000 text characters
Character rate
$0.019 per 1,000
Direction allowance
1,000 unbilled characters
Expression
Write the reply and its mood separatelyA reassuring answer and an excited announcement can use the same preset with different instructions. The direction field accepts emotion, persona, tone, pace and dialect without adding those words to the spoken reply.
Cast
597 presets for character auditionsThe selection includes eight English voices alongside 589 Chinese-language choices. Start with the default Luna Wang or audition another preset from this endpoint before building a character around it.
Control
Shape a line at two levelsUse written direction for the intended performance and numeric speed, pitch and volume for delivery adjustments. SSML can add pauses, pronunciation guidance and emphasis when enable_ssml is on.

What the Qwen-Audio 3.0 TTS Plus API does well

  • Social expression as the starting point

    The Qwen team positions 3.0 Plus for expressive social and companion voices. Try the same clean reply with an encouraging mood and a more matter-of-fact reading to decide what suits the character.

  • A broad Chinese-language cast

    The 589 Chinese-language presets offer a large audition pool for character-led projects. Voice IDs are specific to this model, so take the selection from its own list.

  • Eight English speaker choices

    English presets include Olivia Lin, Luna Wang, Nora Hu, Ivy Hu, Alexander Hu, Adrian Gao, James Zhao and Ryan Ma. Select a voice before refining the emotional direction of a line.

  • Directions that stay off the soundtrack

    A performance note may contain up to 1,000 characters, separate from the 10,000-character spoken script. It is not voiced and contributes no billable characters.

  • Delivery settings for a scene

    Speed and pitch each span 0.5–2, while volume spans 0–100. Save the chosen settings with the character brief, then check subsequent takes for delivery differences.

  • An export for playback or editing

    Receive MP3 by default or request WAV, raw PCM or Opus at a supported rate from 8 to 48 kHz. bitrate applies only when the selected encoding is Opus.

Who Qwen-Audio 3.0 TTS Plus is for

  • A chat character with a recognisable role

    Companion and chat app builders

    A chat character with a recognisable role

    Cast a supportive companion, then describe how that character responds to a small success or a difficult day. Keep the actual reply in text and the emotional context in instructions.

  • A short voiceover with a clear attitude

    Social video creators

    A short voiceover with a clear attitude

    Give a quick update, daily tip or product introduction a conversational read. Compare pacing against the edit before choosing the final voice track.

  • Build a cast for branching scenes

    Interactive fiction teams

    Build a cast for branching scenes

    Assign presets to characters in a visual novel or interactive story. Keep a voice-and-direction note for each role and audition lines from different branches together.

  • Role-play an everyday exchange

    Language-practice app teams

    Role-play an everyday exchange

    Create a practice partner for a cafe order, an introduction or a workplace conversation. Set a language hint and adjust pace so the learner can follow the exchange.

How to call the Qwen-Audio 3.0 TTS Plus API

  1. Write one character response

    Place the exact dialogue in text, with a maximum of 10,000 characters. Begin with a short line that shows the character at their most typical.

  2. Audition the preset list

    Select this model's voice_id or start with the English default, Luna Wang. A preset is the available casting mechanism; there is no cloning input.

  3. Give the character a delivery note

    Describe emotion, tone and persona in instructions without exceeding 1,000 characters. Choose auto detection or one of the ten language hints and tune the numeric controls if needed.

  4. Generate a take and listen in context

    Submit the configured request and poll its task reference for completion. Play the resulting line alongside neighbouring dialogue and download selected files before the 14-day retention period ends.

Qwen-Audio 3.0 TTS Plus worked examples

  • An encouraging chat reply

    The words offer everyday encouragement, while the note establishes a character and an upbeat manner.

    {"text": "You finished the first lesson. That is a good start. Shall we try the next exercise together?", "voice_id": "qwen-audio-3.0-tts-plus-loonglunawang", "language": "en", "instructions": "An encouraging study companion, cheerful without rushing the final question."}
  • A social video about a balcony garden

    A brief voiceover gets a conversational direction and a modest pace adjustment for the edit.

    {"text": "A sunny corner and a few small pots are enough to start a balcony garden. Today, we are planting mint, basil and parsley.", "voice_id": "qwen-audio-3.0-tts-plus-loonglunawang", "instructions": "A presenter sharing a practical idea with relaxed enthusiasm.", "speed": 1.05, "output_format": "wav"}
  • A cafe role-play partner

    The language hint and slower speed prepare a straightforward practice line rather than a complex acting brief.

    {"text": "Good morning. Would you like a table by the window, or would you prefer to sit outside?", "voice_id": "qwen-audio-3.0-tts-plus-loonglunawang", "language": "en", "instructions": "A cafe host speaking patiently to a language learner.", "speed": 0.85}

Qwen-Audio 3.0 TTS Plus API questions

Choose a short exchange that reveals the character’s usual attitude, then audition the voice before adding an elaborate performance note. Change the mood of that same exchange to see whether the preset suits the range of scenes you need. Keep instructions focused on the situation and how the character addresses the listener. Once a voice fits, record its model-specific ID and the settings you selected, while continuing to review each generated line. For interactive stories, listen across branches; for social videos, judge the read against the picture and captions. Both checks are more useful than approving an isolated line.