Skip to main content

Qwen-Audio 3.1 TTS Flash accepts a script and performance directions for multilingual speech, using 70 system voices with configurable speed, pitch, volume and audio output.

From $0.0029/1,000 characters

Workbench

Input

Text to speech
162/10000
3.1 TTS Flash 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

162 characters

3.1 TTS Flash casting: choose one of 70 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.
66/1000
3.1 TTS Flash 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?

Take the same project into more languages

Qwen-Audio 3.1 TTS Flash combines the broadest language coverage of these four Alibaba models with low-latency positioning. Its 70 system voices include four multilingual presets, 15 premium English choices and 51 other-language voices. At $0.00285 per 1,000 text characters, it is the least expensive of these four SpicyAPI endpoints. SpicyAPI layers on no content filter, and this model read the NSFW scripts as submitted in our tests dated November 5, 2026.

Open the playground
Take the same project into more languages
System voices
70 in total
Voice groups
4 multilingual + 15 English + 51 others
Language setting
Auto or 10 explicit hints
Script allowance
Up to 10,000 characters
Price per 1,000 characters
$0.00285
Stored outputs
Available for 14 days
Language breadth
Look beyond the ten hint valuesThe voice catalog covers languages beyond the language enum, including Dutch and Greek. Treat the hint list and the voice selection as different controls, and audition the target-language script with a suitable preset.
Voice groups
70 presets for a multilingual projectChoose among four multilingual voices, 15 premium English presets and 51 voices for other languages. Emily is the default, while the multilingual cast offers another starting point for projects spanning markets.
Budget
The lowest character rate of these four3.1 Flash costs $0.00285 per 1,000 characters in text, calculated per character. The rate excludes instructions, with no charge for failed requests and no extra fee for choosing a different export format.

What the Qwen-Audio 3.1 TTS Flash API does well

  • The widest language remit in this group

    Alibaba positions 3.1 Flash for broad language coverage and low latency. Use that remit to shortlist it for localisation, then check the reading in each language you actually need.

  • A multilingual and language-specific cast

    Four multilingual presets sit beside 15 premium English choices and 51 other-language voices. Select a preset from this model’s voice_id list rather than carrying an ID over from another endpoint.

  • Directions for tone and local delivery

    Describe emotion, persona, pace, tone or dialect in instructions, with room for up to 1,000 characters. This guidance is separate from both the spoken script and the billable text count.

  • Script space for tours and lessons

    A request accommodates up to 10,000 text characters. Divide longer material into reviewable sections and verify the audio before assembling a complete tour or course.

  • Pronunciation and pacing options

    Enable SSML for markup-based pauses, emphasis and pronunciation guidance. Speed and pitch accept values from 0.5 to 2, and volume can be set from 0 to 100.

  • Four encodings at supported sample rates

    Choose MP3, WAV, raw PCM or Opus, with discrete sample rates between 8 and 48 kHz. MP3 at 24 kHz is the default; the bitrate setting concerns Opus only.

Who Qwen-Audio 3.1 TTS Flash is for

  • Narrate a tour in several languages

    Product localisation teams

    Narrate a tour in several languages

    Prepare the translated tour scripts, then audition suitable voices for each language. Review navigation labels and product names before putting the tracks into the product.

  • Give each lesson a spoken version

    Language-learning publishers

    Give each lesson a spoken version

    Turn prepared lesson lines and everyday exchanges into audio. Adjust the reading pace for the exercise and have pronunciation checked before distributing the lesson.

  • Make announcements for visiting guests

    Travel and hospitality operators

    Make announcements for visiting guests

    Create reception information, meeting-point notices and travel announcements in the languages guests need. Listen carefully to place names, times and the instruction at the end.

  • Localise the voiceover for a product video

    Cross-border commerce teams

    Localise the voiceover for a product video

    Record prepared market-specific scripts for demonstrations and product explainers. Judge each take against its own video edit and have the wording reviewed for the intended audience.

How to call the Qwen-Audio 3.1 TTS Flash API

  1. Prepare the target-language script

    Supply the wording in text, with no more than 10,000 characters. This is speech synthesis, so prepare the translated copy before sending it for a reading.

  2. Select a suitable system voice

    Audition the relevant presets in the 70-voice catalog; Emily is used when voice_id is omitted. For a language outside the hint enum, begin with auto and review the reading with an appropriate voice.

  3. Set pronunciation and delivery direction

    Use instructions for the intended tone or dialect, keeping it within 1,000 characters. Enable SSML if needed and tune speed, pitch or volume while comparing samples.

  4. Request audio, then review locally

    Submit with your chosen format and sample rate, and poll the task for its result. Have the target-language performance checked and save the selected file within 14 days.

Qwen-Audio 3.1 TTS Flash worked examples

  • An English product-tour introduction

    Emily reads a short interface tour with measured direction; the voiceover describes the product without adding technical implementation detail.

    {"text": "Welcome to your workspace. Start with a new project, add your notes, and invite your team when you are ready to share.", "voice_id": "Emily_v3.1", "language": "en", "instructions": "A product guide, welcoming and clear, with a deliberate pause between actions.", "output_format": "wav"}
  • A language lesson at a slower pace

    A multilingual preset and a restrained speed setting prepare an everyday English line for a listening exercise.

    {"text": "The library is opposite the station. Walk past the cafe, then turn left at the next crossing.", "voice_id": "longanlingxin_v3.1", "language": "en", "instructions": "A language teacher giving a patient example of everyday directions.", "speed": 0.8}
  • A hotel announcement for guests

    A different English preset and a hospitality brief demonstrate casting a practical announcement.

    {"text": "The morning walking tour meets in the lobby at nine thirty. Please arrive ten minutes early and bring a bottle of water.", "voice_id": "Luna_v3.1", "language": "en", "instructions": "A hotel host addressing a small group, friendly and easy to follow.", "output_format": "mp3"}

Qwen-Audio 3.1 TTS Flash API questions

Organise the project by target language before selecting voices. The ten explicit hints are a request control rather than a complete list of the catalog’s languages, so audition the appropriate preset on the actual translated wording. Include a product name, a number and a full instruction in the first sample; those details often determine whether the recording is usable. Have each target-language reading reviewed before generating the rest of a tour or lesson. Keep the chosen voice ID, instruction note and export settings beside the script, and download approved tracks so the project does not depend on the temporary output storage.