mix-and-master
Skill11 starsChanged 36 days agoMITView on GitHub
Mixing and release workflow in Ableton Live with the ableton MCP tools. It diagnoses a section with mix_doctor, fixes masking and balance, matches a reference track, checks loudness and true peak for streaming platforms, and A/Bs with snapshots. Use when the user asks about their mix, levels, loudness, LUFS, EQ, masking, muddiness, a reference track, or getting a song ready for release.
In this repository5 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- ableton-liveSkill
- ableton-setupSkill
- compose-in-liveSkill
- turkish-makamSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 46
Run mix_doctor again on the same section and report what changed.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Mixing and mastering in Live
Diagnose by listening, change one thing at a time, and let the user compare.
1. Pick the section
Agree on what to listen to: an Arrangement range (start_beat,
length_beats, e.g. the drop) or a Session scene (scene=0). The busiest
8–16 bars are the most useful. Every analysis records in real time.
2. Save the starting point
snapshot(action="save", name="before mix"). It stores mixer and device
settings, so every later change can be compared and undone as a whole.
3. Diagnose
mix_doctor(start_beat=..., length_beats=..., platform="spotify"), or
mix_doctor(scene=0). It records every track and the master in one pass and
reports:
- the master's loudness, true peak, crest factor and tonal balance against the platform's target;
- masking: which two parts collide in which band, and how often;
- per track: stray low end, phase problems, overs, level, main band.
Explain the findings in plain words, most audible first, and propose concrete moves before making them.
4. Fix
- Balance and space:
set_track_mixer(track=..., volume_db=..., pan=..., sends_db={...}). - EQ and dynamics:
add_device(track=..., device="EQ Eight", settings={...}), thenset_device_parameter_by_name(...)to adjust. Live shows parameter names like "1 Frequency A", "1 Gain A". - Ready-made starting points:
apply_device_chain(track=..., chain="drum_bus")(vocal, drum_bus, bass, pad, master, sidechain_duck withsidechain_from). - Masking: cut the less important part 2–4 dB in the shared band, or duck it from the other (sidechain_duck), rather than boosting the other part.
Run mix_doctor again on the same section and report what changed. The user
can A/B with snapshot(action="restore", name="before mix") and a second
snapshot of the new state.
Watching the mix while it plays
With Max for Live available (Suite, or the Trial), install_analyzer() puts
the AbletonMCP Analyzer on the master (or track=...), and
read_analyzer(seconds=4) reports octave bands, loudness, correlation and
width per analyzed track while the set plays, without recording. Use it to
check a change quickly; use mix_doctor and render_and_analyze for the
full, exact picture.
5. Match a reference
compare_to_reference(reference_path="/path/to/reference.wav") compares
loudness, true peak, dynamics, width and 1/3-octave tonal balance with levels
matched, using the reference's loudest stretch. It also returns
eq_eight_settings that move the mix about 60% of the way. Apply them only if
the user agrees: add_device(track="master", device="EQ Eight", settings=...).
6. Release check
Targets (integrated loudness / true peak): Spotify, YouTube, Tidal,
SoundCloud -14 LUFS / -1 dBTP; Apple Music -16 LUFS; Amazon Music -14 LUFS /
-2 dBTP; club masters often -6 to -9 LUFS / -0.3 dBTP. Louder than the target
only gets turned down by the platform, at the cost of punch. Measure the whole
song with render_and_analyze(start_beat=0, length_beats=...); for the
master chain, apply_device_chain(track="master", chain="master") ends in a
true-peak limiter at -1 dB.
The MCP prompts mix_check, match_reference and release_check run these
workflows step by step.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

