Complete guide to using A Complete Audio Generator for Speech and Music
PoYo Audio Generator brings leading speech, voice, and music generation models into one focused workspace. Instead of switching between separate provider tools, you can choose ElevenLabs, Gemini, xAI, MiniMax, and other executable audio models directly inside the Input panel.
Each model keeps its native controls, so language, voice, emotion tags, speed, lyrics, instrumental mode, output format, preview, JSON response, and download workflows appear only when the selected model supports them.
Generate speech, voiceovers, and music, then preview, download, and move stable workflows into your API pipeline.
01
Turn scripts, product copy, narration, and dialogue into natural audio. Supported models expose voice selection, language control, emotion tags, text normalization, speed, and timestamps when available.
02
Create music from prompts, lyrics, instrumental settings, and structured composition ideas. Use text-to-music models for demos, hooks, backing tracks, and creative direction before moving into production.
03
Switch between audio model families without losing their native controls. Fast TTS, expressive voice models, music generators, sample rate, bitrate, output format, and quality options stay attached to the active model.
04
Listen to generated audio directly in the browser, inspect structured JSON and timestamps when returned, download the final file, and move validated settings into API or agent workflows.
Create narration, product walkthrough audio, multilingual dubbing drafts, and social video voice tracks from scripts.
Generate spoken lessons, chapter previews, podcast segments, and long-form narration tests before recording or mastering.
Prototype app voices, screen-reader style audio, onboarding narration, and localized speech experiences.
Create song ideas, instrumental beds, hooks, demos, and backing tracks from prompts, lyrics, and style direction.
Explore character voices, game UI sounds, creator assets, and audio concepts for interactive media.
Validate prompts, voices, formats, previews, and JSON responses before adding audio generation to an API product.
1. Choose an audio model in the Input panel. Pick the provider and executable model that matches your speech or music workflow.
2. Add text, lyrics, or music direction. Write the script, prompt, language, voice style, mood, instrumentation, or lyrics required by the selected model.
3. Adjust model-specific controls. Set only the options shown by the active model, such as voice, emotion tags, speed, sample rate, bitrate, output format, lyrics optimizer, or instrumental mode.
4. Run, preview, and download. Submit the generation, listen to the audio preview, inspect JSON if needed, and download the generated file.
PoYo Audio Generator can create text-to-speech, voiceover audio, narration, dialogue, music demos, songs, and instrumental tracks depending on the selected model. The Input panel only shows the fields supported by the active audio model.
Start with ElevenLabs V3 TTS for expressive voiceovers, ElevenLabs TTS Turbo v2.5 or xAI TTS 1 for faster speech tests, Gemini 3.1 Flash TTS for style-instructed speech, ElevenLabs Music for structured music generation, MiniMax Music 2.6 for prompt and lyrics workflows, and Generate Music for broad music creation.
Quality depends on the model, prompt, language, voice, and output settings. Modern TTS models can sound highly natural, but production use still benefits from listening checks, pronunciation edits, pacing review, and final mastering.
Yes, when the selected model supports those controls. Some speech models expose language, voice, speed, stability, style, emotion tags, or timestamps. Music models may expose genre, mood, lyrics, instrumental mode, section structure, and output format.
The gallery previews common audio files in the browser. Depending on the provider response, downloads can include MP3, WAV, OGG, Opus, PCM, mu-law, A-law, timestamps, or supporting files.
Use clean scripts, clear pronunciation hints, specific voice direction, realistic pacing, and concise emotion tags. For music, describe genre, mood, tempo, instruments, vocals, structure, lyrics, and what to avoid.
Generation time and cost vary by model, text length, duration, format, music length, quality settings, and provider. Check the active model cost in the Input footer before running longer narration or music generations.
Commercial use depends on the selected model, provider terms, your account plan, and the rights attached to scripts, lyrics, voices, references, and uploaded material. Review the relevant terms and avoid content you do not have permission to use.
Compare speech, voice, and music generators without rebuilding prompts across provider tools.
Each model keeps its own voice, language, style, lyrics, format, and result workflow.
Listen to generated audio directly in the output gallery before downloading files.
Test scripts, prompts, voices, formats, and JSON responses before moving into API-powered products.
