skills/spatial9-remix-by-voice/SKILL.md

SKILL.md: full Markdown source from the spatial9-remix-by-voice folder of Spatial9 Agent Skills. Free, MIT licensed code for immersive audio with the...

Raw file · View on GitHub · Download every file as a zip

Spatial9 remix by voice

Requires the Spatial9 MCP server at https://spatial9.ai/mcp and the job_id of a finished render.

Steps

  1. Read the current mix with get_mix (stems, positions, levels) and get_effects. For song structure, call analyze_song.
  2. Translate the request into remix_render changes:
  • Position: direction and height, or exact azimuth_deg, elevation_deg and distance.
  • Level: gain_db (absolute) or gain_change_db (relative), or mute.
  • Motion: motion.shape, one of circle, spiral_up, rise, sink, sweep, fly_over, come_close, move_away, wander or none. Add when (for example chorus), speed and clockwise.
  1. Always run remix_render with preview: true first. It is free. Describe what will change and ask the user to confirm.
  2. After confirmation, run the same changes without preview and with an idempotency_key.

Effects and spaces

  • find_effects with a query (over 1,400 presets). Then apply_effects with action: add and a preset_id, or up to 6 effects with their params.
  • find_spaces for rooms, halls and venues. Apply the best match with apply_effects.
  • Use target or stem_id to choose which stem, or the master bus, the effect goes on.

Rules

  • Keep lead vocal, kick, snare and bass near the front centre unless the user insists otherwise.
  • Every movement needs a musical moment. Avoid constant motion; it causes listener fatigue.
  • Save a look the user likes with save_mix_template so it can be reused with template_id.