Mirelo AI has deployed its latest tool for the creative stack: Sound Agent. This new AI agent fundamentally shifts how video sound design is handled, moving away from manual, track-by-track timeline manipulation to a prompt-driven interface. Users simply type commands like "Sound design this clip," and the system handles the heavy lifting of audio separation and editing.

Under The Hood

The core functionality relies on advanced audio source separation. Upon receiving a prompt, Sound Agent splits a video’s audio into distinct stems: speech, sound effects, and music. It then edits these layers independently based on the user's natural language instructions. This allows for granular control without the need for traditional digital audio workstation (DAW) expertise.

Prompt-Driven Capabilities

The agent supports a wide range of specific audio tasks via simple text commands. It can generate new sound effects, remove unwanted background music while preserving effects, or strip soundtracks entirely. For dialogue-heavy content, it offers dubbing capabilities into English, Spanish, French, and Mandarin, as well as splitting speech by individual speaker. Precision is also possible; users can request specific sounds at exact timestamps, such as "At 2.4 s, add a heavy thud," or dynamically change a score's mood from calm to intense.

Reference Inputs and Pricing

Beyond text prompts, the agent accepts reference audio, allowing users to hum a melody or upload a sound file to guide generation. This multimodal input approach reduces the friction of describing abstract audio concepts in words. Pricing is structured for various scales, starting at $20 per month for the Creator tier and $60 for Studio. A pay-as-you-go option is available at $1.50 per video, with credit costs varying by frame rate: $0.60 per minute for 24 fps content, $0.75 per minute for 30 fps, and $1.50 per minute for 60 fps.

Why It Matters

For content creators and editors, this tool targets one of the most time-consuming aspects of post-production. By automating the separation and manipulation of audio stems, Mirelo AI aims to drastically reduce the turnaround time for short-form video, where audio quality often dictates viewer retention. It represents a clear trend of AI agents absorbing complex, specialized workflows that previously required significant manual effort.

Key Takeaways

  • Sound Agent uses natural language prompts to edit video audio stems (speech, SFX, music).
  • Supports dialogue dubbing in English, Spanish, French, and Mandarin.
  • Pricing includes a $1.50 per-video credit or monthly subscriptions starting at $20.
  • Accepts reference audio inputs like humming or uploaded files to guide sound generation.

The Bottom Line

Sound design is the final frontier of automation in video editing, and Mirelo's agent proves that natural language is finally good enough to replace the timeline. This isn't just a convenience; it's a fundamental shift in who gets to be a creator, putting studio-quality audio in the hands of anyone with a prompt.