Your first session with Maestra: ten things to say

The short answer: start with small, direct requests on a set you know well. Ask Maestra to read the session, then mute, level, rename and launch things. Then move on to devices and plug-in parameters. Each request below is something Maestra handles today. Each one makes a change you can check in Live straight away.

You need Maestra installed and connected to Ableton Live 11.3 or later, any edition. Open a copy of a real project, or a disposable one, before you begin. Every change goes through Live's own undo, so "undo that" walks back the last one.

How to read the replies

Maestra reads the session, makes the change and then reads the result back from Live. Its reply reports what the read-back confirmed. If a change was sent but did not verify, it says so plainly.

Track and device names are matched loosely. "the bass" finds a track called Bass. If a name could mean two things, Maestra asks which one you meant.

1. "What's in this session?"

This is a read. Maestra lists the tracks, devices and scenes it sees and changes nothing. It is the best first message, because every later request is grounded in what this scan found.

2. "Mute the bass."

A direct mixer command. "Unmute the bass", "solo the vocal" and "arm the Keys track" work the same way. With Turbo on, a quick decision step picks the action for direct commands like this one, so they skip the full model turn.

3. "Set the Drums volume to 70%."

Maestra sets the track's mixer volume and reads it back. You can also say "turn the Drums down a bit" and let it pick a sensible step. Pan and sends work the same way: "pan the hats left" or "set the vocal's send A to 30%".

4. "Rename 3-Audio to Lead Vox."

Renaming is live, with no reload. This one is useful on inherited projects full of default names. For a whole set, ask for a plan first: "suggest names for every track", then approve it.

5. "Fire scene 2."

Scenes are numbered as Live shows them. Scene 1 is the top row. "Add a scene" appends an empty one at the end. "Duplicate scene 3" copies a whole row.

6. "Launch the clip on Drums in scene 3."

Maestra first checks the slot actually holds a clip, because firing an empty slot does nothing in Live. Then it reads back that the clip is playing, or triggered to start at the next quantization boundary, before it reports success.

7. "Put a Compressor on 3-Audio."

Live's own devices load straight onto the track on Live 12.3 or later. If your Live is older, Maestra names the version the tool needs and the one you are running.

Plug-ins and your own Max for Live devices load too: "load Pro-Q 4 on the vocal". Maestra searches the plug-ins in Live's browser first. If it finds more than one match, or none, it shows you what it found and lets you pick. It never swaps in a plausible-sounding alternative on its own.

8. "What can I control on Serum?"

Maestra lists every parameter name the plug-in exposes to Live. It also marks which of them are set up for control in Live. This is a read, so nothing changes. The list matters because Live only reaches the parameters in its device panel. Our plug-in guide explains why.

9. "Set Serum's Fil Cutoff to 40%."

Maestra changes plug-in parameters by name. On a control that shows no unit, a percentage sets the knob position, and the reply says so. On a control that shows dB, Hz or ms, give the value in that unit: "set the attack to 250 ms".

If the parameter is not set up in Live yet, Maestra leaves every other control alone. On a VST2 plug-in on Windows it can link the parameter itself; otherwise it tells you how to add it with Configure and then rescans.

10. "How does Turbo work?"

Ask Maestra how Maestra works. It answers in the chat and changes nothing in your set. Good questions to try:

• "Where is voice input?"
• "What does Turbo cost?"
• "How do I set up Quick Chat?"
• "Live isn't connecting. What should I try?"

It tells you where each control is, and you make the change in Settings yourself.

What happens with bigger requests

Requests that touch several tracks, like "set up this session for mixing", go through a plan. Maestra lays out what it will do per track. How far it goes before asking depends on the autonomy setting in the row under the chat box:

• Careful asks before anything unclear and shows the plan first. Every change waits for your ok.
• Balanced acts when the request is clear and tells you what it assumed. This is the default.
• Yolo acts on its best reading and reports after.

On every setting, deletes and other destructive changes wait for your ok. Ctrl+Shift+A (Cmd+Shift+A on a Mac) cycles the three.

Credits in a first session

Chat messages use credits. A credit is one exchange with the model. A simple request usually costs one or two. A bigger build costs more, because Maestra works through it in steps. Turns that go through Turbo's decision step use 1.5× credits. Your balance is in Settings › Activation.

Summary

Start with a read, then climb: mute, level, rename, fire scenes and clips, load devices, then list and set plug-in parameters by name. Each change is read back from Live, and "undo that" walks it back. Ask Maestra how Maestra works whenever you are unsure. To try the ten on your own session, start on the Get started page.