Speech 2.8 HD
Native sound tags add breaths, laughs, and natural pauses across 40 languages.
Try Speech 2.8 HD
Paste a script, pick a voice, and generate.
Professional voiceover for every use case
Audiobooks and long-form reading.
Turn chapters into steady, expressive narration while preserving pauses and shifts in mood.E-learning and course modules.
Deliver instructions and definitions at a measured pace that helps learners follow each step.Explainer and product demos.
Give product walkthroughs a polished voice that keeps features and instructions easy to understand.IVR and phone prompts.
Generate calm, consistent menu prompts for support lines, appointment systems, and business phone routing.Adds breaths, laughs, and pauses to speech
Speech 2.8 HD supports native sound tags for audible actions such as breathing, laughing, sighing, and clearing the throat. These details help conversational scripts avoid the flat rhythm associated with basic text-to-speech.
- •Native interjection and sound tags
- •Breaths, laughs, sighs, and gasps
- •Seven supported emotion modes
- •Custom pause markers within scripts
- •HD output with reduced background artifacts
Choose from 300+ system voices
MiniMax provides more than 300 system voices alongside support for 40 languages and selected dialect enhancements. This range makes it easier to match narration with the audience, language, and intended destination.
- •300+ provider system voices
- •40 supported languages
- •Different ages, tones, and delivery styles
- •Language detection and language boosting
- •Support for selected accents and dialects
Clone an authorized voice from a short sample
Speech 2.8 HD can create a custom voice from a short reference recording, with MiniMax recommending about ten seconds of clear speech. This can keep approved brand, creator, or character narration consistent across projects.
- •Cloning from a short reference recording
- •Mono and stereo source audio supported
- •Captures timbre and speaking pace
- •Works with synchronous and long-text synthesis
- •Requires ownership or explicit permission
Controls pacing, pronunciation, and audio delivery
Adjust speed, pitch, volume, pauses, pronunciation, sample rate, bitrate, and file format for different production needs. Synchronous requests support shorter scripts, while asynchronous generation handles long-form projects.
- •Speed range from 0.5 to 2.0
- •Pitch and volume adjustment
- •Inline pronunciation guidance
- •Sentence-level and word-level timestamps
- •MP3, PCM, FLAC, and WAV options
- •Long-text asynchronous generation
How it works
1
Paste your script
Enter the exact words you want the voice to read. Use punctuation and paragraph breaks to organize the delivery, and add supported sound or pause tags when the script needs extra expression.
2
Pick a voice
Choose a voice that suits your audience and destination, such as a calm course narrator or an upbeat product presenter. Check the language and listen to a short preview before generating the full script.
3
Generate and download
Create the voiceover and listen for pacing, pronunciation, and tone. Revise the script or voice settings if needed, then download the finished audio for editing or publication.
Pricing for Speech 2.8 HD
Runs on credits — no per-model surcharges, no surprise billing.
15credits
per 100 characters
≈ 90 credits for a 600-character script (about a minute of speech)
Frequently asked questions
What is Speech 2.8 HD?▾
Speech 2.8 HD is MiniMax's high-quality text-to-speech model for turning written scripts into generated voice audio. Its main capabilities include native sound tags, 40-language support, a large system voice library, and voice cloning.
How does Speech 2.8 HD work?▾
You paste a script, choose an available voice, and generate the reading. The model interprets punctuation, paragraph breaks, pause markers, language settings, and supported sound tags to shape the delivery.
What languages does Speech 2.8 HD support?▾
Speech 2.8 HD supports 40 languages, including English, Chinese, Cantonese, Spanish, French, German, Portuguese, Arabic, Japanese, Korean, Hindi, and many others. Language detection and targeted language boosting are also available.
How many voices does Speech 2.8 HD have?▾
MiniMax lists more than 300 system voices for synchronous text-to-speech, covering varied tones, ages, and speaking styles. The model can also use compatible custom cloned voices.
Can I change the emotion, speed, or pronunciation?▾
The model supports seven emotion modes as well as controls for speed, pitch, volume, and scripted pauses. Pronunciation can be guided with a dictionary or inline phonetic notation where the interface supports those settings.
Can I clone my own voice with Speech 2.8 HD?▾
Yes, MiniMax supports voice cloning from a short, clear reference recording and recommends about ten seconds of audio. You may only clone a voice you own or have explicit written permission to use.
How long can a Speech 2.8 HD script be?▾
Synchronous text-to-speech requests must contain fewer than 10,000 characters. MiniMax also provides an asynchronous long-text workflow that can process up to one million characters per request, including uploaded text files.
What audio formats does Speech 2.8 HD output?▾
Supported formats include MP3, PCM, FLAC, and WAV, although WAV is limited to non-streaming generation. Available settings include mono or stereo output, multiple bitrates, and sample rates up to 44.1 kHz.
Can I use Speech 2.8 HD voiceovers commercially?▾
The model's technical capabilities do not automatically grant commercial rights. Commercial use depends on the current service terms and your rights to the script, selected voice, and any cloned recording, so review the applicable terms before publishing.
How much does Speech 2.8 HD cost?▾
On BudgetPixel, speech generation is billed per 100 characters of script. Longer scripts use more credits, so the total cost scales with the amount of text rather than the number of audio files.
