Aether

Describe the part. Get the MIDI.

Aether is an AI co-producer that lives on a track in your session. Instead of clicking every note into a piano roll, you tell it what you want and it plans the part, generates the MIDI, and hands you notes you can drag straight onto a track:

• "A dark, driving minor-key bassline at 124 BPM"
• "A warm lo-fi chord progression in F minor"
• "Add a horns section under the piano"

Aether reasons about key, scale, voice leading, and rhythm, so the parts it writes sit in your track instead of fighting it. Don't like something? Ask for a change in the next message and it refines the MIDI it already wrote rather than starting over, so you iterate toward the idea in your head. Everything Aether produces is standard, editable MIDI.

New in 1.2: compose in layers

Arrangements are built from any number of named, instrument-tagged layers, not a fixed set of roles. Ask for a new part and it lands alongside what you already have instead of replacing it:

• Additive by design: "add an ostinato under the piano" creates a new layer; your existing layers stay exactly as they were.
• Layers view: every layer listed with its instrument, a mini piano-roll preview, and per-layer solo and mute.
• Color-coded piano roll: each layer gets its own color, so overlapping parts stay readable.
• Named multi-track export: one named track per layer, stamped with its General MIDI program, with tempo and time signature intact. Drag the whole arrangement to your DAW, or a single layer on its own.

Key features:

• Text-to-MIDI: melodies, chord progressions, basslines, and full layered arrangements from natural-language prompts.
• Conversational refinement: follow-up messages edit the MIDI it already wrote.
• Music-theory-aware phrasing: key, scale, voice leading, and rhythm.
• Drag-to-DAW workflow: generated MIDI drags straight onto any track; Lock to Host syncs BPM and time signature from your DAW.
• MIDI Library: every generation auto-saves with idea and version history.
• Producer tools built in: dual-mode piano roll, post-processing (Raw / Clean / Enhance), live MIDI capture, and reusable custom prompts.

Bring your own AI

Connect the model you already pay for with your own API key: Claude (Anthropic), ChatGPT (OpenAI), Gemini (Google), Grok (xAI), DeepSeek, OpenRouter, or Hugging Face. Or run a local model through Ollama or LM Studio and generate completely offline. No forced subscription, ever.

Formats and system requirements:

• Formats: VST3, Audio Unit (AU), and Standalone.
• OS: macOS and Windows.
• Hosts: Ableton Live, FL Studio, Logic Pro, Cubase, Studio One, Bitwig, Reaper, and other AU/VST3 hosts, or standalone with no DAW.
• Requires: an AI provider API key, or a local model via Ollama / LM Studio. Internet connection required for cloud providers only.

Pricing and licensing

One-time purchase of $60. Free lifetime updates across Aether 1.x, activation on up to three machines, and no subscription required when you use your own API keys.