Generate music tracks from text
# ElevenLabs Music High-fidelity AI music generation for professional creators. ## Overview ElevenLabs Music is a state-of-the-art generative audio model designed to transform text prompts into high-quality, expressive musical compositions. By leveraging advanced deep learning architectures, the model interprets complex descriptive language to synthesize instrumental tracks, melodic arrangements, and rhythmic soundscapes. It is engineered to bridge the gap between creative intent and technical execution, allowing users to generate professional-grade audio assets without requiring traditional music production software or extensive sound engineering knowledge. What makes this model distinct is its nuanced understanding of musical structure and emotional tone. Unlike generic audio generators, ElevenLabs Music excels at maintaining coherence over longer durations, ensuring that transitions between sections feel natural and intentional. Whether you are looking for a subtle background bed for a social media ad or a high-energy beat for a brand campaign, the model provides a versatile toolkit for rapid prototyping and final-asset creation. ## Use Cases * Generating custom background music (BGM) for short-form video ads on TikTok and Instagram. * Creating unique, royalty-free soundtracks for corporate brand videos and presentations. * Developing rhythmic audio beds for podcast intros and outros to enhance brand identity. * Producing atmospheric soundscapes for immersive digital experiences and interactive web ads. * Rapidly iterating on musical concepts during the pre-production phase of ad creative development. * Customizing instrumental tracks to match the specific pacing and mood of a video edit. ## Parameters This model offers a robust set of configuration options to fine-tune your audio output. Please refer to the Input Parameters section in the Pixloop interface to adjust settings such as text prompts, output formats, total duration, and instrumental-only constraints to achieve the desired audio profile. ## Tips for Best Results * Be descriptive with your prompts: Include genre, tempo, mood, and instrumentation (e.g., "Upbeat lo-fi hip hop with a mellow piano melody and crisp drums"). * Utilize the 'force_instrumental' toggle if you are creating background music for voiceovers to ensure the AI does not introduce competing vocal tracks. * Experiment with the 'sections' parameter to create dynamic shifts in your music, such as building tension or introducing a bridge. * Start with a 30-second duration for initial testing to iterate quickly on the musical style before generating longer, full-length tracks. * Adjust the 'output_format' based on your final delivery platform; use higher bitrates for high-fidelity video projects and lower bitrates for web-optimized mobile ads. * If the output feels too chaotic, simplify your prompt by focusing on one or two core musical elements rather than layering too many conflicting descriptors. ## About ElevenLabs Music is developed by ElevenLabs, a leader in generative AI audio technology. Building upon their expertise in high-fidelity voice synthesis, this music model represents a significant expansion into the broader audio generation landscape. The model is optimized for creative workflows, providing a seamless integration point for marketing teams and ad agencies looking to scale their production of custom, high-impact audio assets.
Model input reference derived from preset schema.
| Name | Type | Required | Default | Description |
|---|---|---|---|---|
| text | string | Yes | - | Text description of the music to generate |
| model_id | select | No | "music_v1" | - |
| sections | array | No | - | Define specific sections of the song with custom styles and lyrics |
| output_format | select | No | "mp3_44100_128" | - |
| music_length_ms | number | No | 30000 | Total length of the music in milliseconds |
| force_instrumental | boolean | No | false | Force the generated music to be instrumental (no vocals) |
| respect_sections_durations | boolean | No | true | Whether to strictly follow the duration specified in each section |