Skip to main content

Qwen-Audio 3.1 TTS Plus accepts a script and performance directions for high-fidelity narration, using 4 multilingual system voices with configurable speed, pitch, volume and audio output.

From $0.029/1,000 characters

Workbench

Input

Text to speech
163/10000
3.1 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

163 characters

Voice
Long An Huan (Chinese)Long An Lingxin (Chinese)Xu Nanchuan (Chinese)Long An Fengyue (Chinese)
3.1 TTS Plus casting: choose one of 4 multilingual system voices belonging to this model; other-model IDs are rejected, no default is specified in the supplied facts, and selection does not affect price.
75/1000
3.1 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?

Cast the narrator. Shape the performance.

Qwen-Audio 3.1 TTS Plus is Alibaba's fidelity-led option for narration, brand recordings and acted dialogue. Choose from four multilingual presets, send a passage of up to 10,000 characters and direct its emotion, persona or pacing in a separate note. The price is $0.0285 per 1,000 text characters, with no charge for directions. SpicyAPI adds no content filter; in our November 5, 2026 tests, this model read NSFW scripts as supplied.

Open the playground
Cast the narrator. Shape the performance.
Cast
4 multilingual presets
Passage length
Up to 10,000 characters
Performance notes
Up to 1,000 characters
Exports
MP3, WAV, PCM, Opus
Narration price
$0.0285 / 1,000 text characters
File retention
14 days
Casting
Four voices to audition for the roleLong An Huan, Long An Lingxin, Long An Fengyue and Xun Anchuan make up the preset cast. Compare an opening passage before choosing the narrator for a book or the speaker for a brand.
Performance
Give the voice a scene to playKeep the spoken passage in text and put the acting brief in instructions. Emotion, tone, persona, pace and dialect belong in that brief, which is neither read aloud nor billed.
Production
A recording format for the next stageMP3 is the default; choose WAV, raw PCM or Opus when the editing or playback workflow calls for it. Available sample rates run from 8 to 48 kHz, with 24 kHz selected by default.

What the Qwen-Audio 3.1 TTS Plus API does well

  • Narration built around fidelity

    Alibaba positions 3.1 Plus for high-fidelity readings, making it a candidate for passages where the narrator carries the work. Audition descriptive prose and dialogue together before committing to a full recording.

  • Room for a substantial passage

    Each task accepts as many as 10,000 characters in text. Divide a longer manuscript at sensible scene or paragraph boundaries, then review the joins between recordings.

  • A separate performance brief

    Use up to 1,000 instruction characters to describe the speaker and emotional arc. For a brand greeting, for example, ask for measured confidence and a welcoming finish.

  • Markup for detailed phrasing

    Switch enable_ssml on when the text contains SSML for pronunciation, emphasis or pauses. Check a marked-up excerpt before applying the same approach across a chapter.

  • Pace, pitch and loudness controls

    Adjust speed and pitch from 0.5 to 2, and volume from 0 to 100. These settings let you audition delivery changes without rewriting the passage.

  • Multilingual casting and dialect direction

    Let language detect the script automatically or supply one of ten language hints. A dialect request can go in instructions; review the pronunciation rather than assuming that direction alone settles it.

Who Qwen-Audio 3.1 TTS Plus is for

  • Give a long passage a narrator

    Audiobook producers

    Give a long passage a narrator

    Start with the opening of a chapter, choose a preset and shape its pace around the prose. Keep a record of that casting choice when preparing later passages.

  • A welcome and menu in the same voice

    Brand and telephone teams

    A welcome and menu in the same voice

    Prepare the greeting and phone-menu wording with a shared preset and brand direction. Listen for clear option names and an ending that leaves the caller time to respond.

  • Audition the performance before the edit

    Film and animation editors

    Audition the performance before the edit

    Generate prepared dialogue with notes about the character and scene. Bring the selected audio into your edit and check its duration against the picture; synthesis does not promise automatic timing alignment.

  • Direct a character through a scene

    Game writers and audio teams

    Direct a character through a scene

    Write the quest line separately from the performance note. Compare a restrained take with a more urgent delivery, then choose the recording that fits the scene.

How to call the Qwen-Audio 3.1 TTS Plus API

  1. Prepare a passage to audition

    Put the words intended for the recording in text, staying within the 10,000-character limit. Use a short excerpt while deciding on the narrator.

  2. Choose a voice from this cast

    Set voice_id to a 3.1 Plus preset; IDs belonging to another model are not accepted. Add language if you want to specify a reading hint rather than use auto.

  3. Describe the reading

    Write the acting direction in instructions, up to 1,000 characters. Set speed, pitch or volume for the take, and enable SSML if your script uses its markup.

  4. Select the export and submit

    Choose mp3, wav, pcm or opus and a supported sample rate. Retain the task reference after submitting, then poll until the recording is ready.

  5. Review and keep the selected take

    Listen for wording, pronunciation and performance before publication. Download the recording within its 14-day storage window.

Qwen-Audio 3.1 TTS Plus worked examples

  • The opening of a bookshop story

    A brief narrator note asks for a measured reading, while WAV prepares the result for an editing workflow.

    {"text": "At six each morning, the bookshop lights came on before the rest of the street. Today, a small blue parcel waited beside the door.", "voice_id": "longanlingxin_v3.1", "language": "en", "instructions": "Read as an observant narrator, with gentle curiosity and an unhurried pace.", "output_format": "wav"}
  • A hotel telephone welcome

    Brand direction stays outside the greeting, so the request separates what callers hear from how it should sound.

    {"text": "Welcome to Cedar House Hotel. For reservations, press one. To speak with our reception team, press two.", "voice_id": "xunanchuan_v3.1", "language": "en", "instructions": "A polished welcome from a host, friendly and composed, with clear menu options.", "speed": 0.95}
  • A game guide at the harbour

    The same clean line can be auditioned with different acting notes before the audio team picks a take.

    {"text": "The ferry leaves at sunrise. Bring your map to the harbour, and I will show you the safest route through the islands.", "voice_id": "longanhuan_v3.1", "language": "en", "instructions": "An experienced guide giving practical advice, quietly confident with a hopeful ending.", "seed": 420}

Qwen-Audio 3.1 TTS Plus API questions

Begin with a passage that contains the kinds of sentences the narrator will meet throughout the work: description, dialogue and a change of mood. Audition the four presets against that passage, then refine the direction for the strongest candidate. Keep the voice, script and seed steady while comparing a single adjustment to pace or performance; a seed is a comparison aid, not a promise of an identical take. For dubbing, review the returned duration in the edit as well as the reading itself. Preserve the selected settings alongside the downloaded audio so later passages have a clear starting point.