# Audio tools for your AI agent


## Local audio tools for agents

Use PAM for local sample generation, WAV loudness and true-peak analysis, model management and persistent audio jobs. Generation and analysis are free. The fx, audio master, preset render and dj set commands process files and render presets and DJ sets without opening PAM; they require a PAM licence or active trial. Inspect the installed model catalogue for runtime availability. To install the command-line tools, run PAM’s macOS installer from https://map.audio/download, click Customize and tick “Command-Line & AI Agent Tools” (Apple Silicon Macs). Per-client file-delivery acceptance remains open. Start with [PAM for agents](/agents), [setup](/agents/setup.md), the [task skill](/.well-known/agent-skills/pam-audio-gen/SKILL.md) and the [installed-command reference](/agents/commands/v1.md). Local shell or MCP execution is required.

# PAM command-line tools: setup and first success

The command-line tools ship inside PAM’s macOS installer, the same download as the PAM trial. There is no separate package, and the CLI has no licence of its own.

To install the command-line tools, run PAM’s macOS installer from https://map.audio/download, click Customize and tick “Command-Line & AI Agent Tools” (Apple Silicon Macs). It installs `pam`, `pamd` and `pamd-worker` under `/Library/Application Support/MAP Audio/PAM/bin` and links `/usr/local/bin/pam`. The choice is unticked by default and independent of the selected plugin formats; run the installer again to add it to an existing PAM. Intel Macs and the Windows installer do not offer it, and no Linux package is distributed.

Use a supported local execution environment. A shell-capable agent runs `pam … --json`; a local MCP client starts `/usr/local/bin/pam mcp`. Ordinary hosted web/mobile chat cannot run your local binary. Per-client installed-package and file-delivery acceptance remains open; no client is claimed release-verified yet.

```json
{"mcpServers":{"pam":{"type":"stdio","command":"/usr/local/bin/pam","args":["mcp"]}}}
```

The `pam` server exposes file and model tools alongside controls for a selected running PAM instance. `pam mcp` relays to MCP sessions hosted by `pamd`; plugin control uses the installed bundled runtime. The `pam-audio-gen` task skill is published at https://map.audio/.well-known/agent-skills/pam-audio-gen/SKILL.md.

Inspect the local contract first:

```sh
pam version --json
pam manifest --json
pam search "measure loudness"
pam schema audio generate --json
```

For a WAV measurement, skip model setup:

```sh
pam audio analyze recording.wav --json
```

For generation, inspect `pam model list --json`. Both command availability and model `runtimeReady` must permit the operation. SA3 and both Magenta models have registered adapters and built-in Apple Silicon profiles. Their model files and native runtime dependencies must still be installed. A downloaded model alone is insufficient. Report `daemonReason` without blind retries.

When the platform supports the model, obtain consent for any missing model download using the client's normal consent flow. SA3 Small SFX weights need about 2.7 GB. Magenta needs Apple Silicon, macOS 14+ and its runtime. The CLI/daemon may build on other platforms; that is not evidence of model eligibility.

```sh
pam model pull stable-audio-3-small-sfx --json
pam audio generate -p "short metallic kick, tight transient" -d 1.2 --category kick --seed 7 -o kick.wav --json
pam audio analyze kick.wav --json
```

Generation, model management, streaming and analysis are free. Check entitlement only if the installed contract marks the command paid. `fx`, `audio master`, `preset render` and `dj set` process files and render presets and DJ sets without opening PAM; they require a PAM licence or active trial. Presets and song references also follow desktop pack ownership: unowned packs return `pack_license_required`, while factory presets need no separate pack purchase. Model licences apply separately: [Stability AI terms](https://stability.ai/community-license-agreement), including commercial registration, organization/affiliate revenue conditions, attribution and separate Gemma encoder terms. Powered by Stability AI.

Wait for the actual job; inspect errors, completion, resulting files and measurements. Deliver the file with the client's supported file mechanism. `wavPath` alone does not establish attachment, upload or audition. Never infer measurements from a prompt.

For musical edits, enable the MCP Bridge in PAM's Settings → Integrations and select the instance in the same `pam` MCP server. Generate and measure files, load them into cells, edit, save, reopen and export through that session. Preserve unsaved work and query applied state after edits; a queued acknowledgement alone does not prove the session changed.

Troubleshooting: missing binary → install the command-line tools as above; unavailable runtime → report readiness; missing model → ask for download consent; `model_in_use` → stop the holder deliberately before retrying; no audible output → inspect the real WAV and measurements. Use `pam daemon status --json`, `pam model ps --json` and job commands for observed state.

## Interfaces

pam provides CLI JSON; pam mcp relays file, model and selected-instance plugin tools through one daemon-hosted MCP server. pamd retains model workers and jobs. Plugin control requires the installed bundled runtime and PAM's existing MCP Bridge toggle. PAM hosts its own inference in-process.

## Recorded demonstration

The page replays three runs recorded on 2026-10-18 with pam 1.8.5 from PAM's installer on an Apple M4 Max. Each made a sample with `pam audio generate`, measured it with `pam audio analyze` and repeated the generation through `pam mcp`. It is recorded playback, not live inference.

## Comparison

Competitor capabilities below come from publisher documentation and source; they were not independently tested. No overall superiority claim is established. Checked 2026-10-18.

| Tool / evidence | CLI and discovery | Generation and analysis | Editable session and verification | Persistent jobs and setup |
| --- | --- | --- | --- | --- |
| [PAM](/agents/commands/v1.md) (pam 1.8.5) | Native CLI JSON, manifest/search/schema and MCP from one command contract. | Local SA3/Magenta generation and Magenta streaming; WAV loudness and true-peak analysis. | One pam MCP server for file/model tools and reading/editing a selected PAM instance. Headless preset render, fx, audio master and dj set require a PAM licence or active trial. | Retained model workers, inspectable jobs and cancellation. To install the command-line tools, run PAM’s macOS installer from https://map.audio/download, click Customize and tick “Command-Line & AI Agent Tools” (Apple Silicon Macs). Inference also needs downloaded weights. |
| [Ableton MCP](https://github.com/ahujasid/ableton-mcp) (1.4.5 source) | MCP with typed tools; session and capability readers. | Creates MIDI content and controls Live devices; no standalone model-job service established by the inspected source. | Track, clip and device edits in running Live; state readers and snapshots. | Requires Live Remote Script and Python 3.10+. Persistent model jobs not established. |
| [Ableton MCP Pro](https://github.com/nicholasbien/ableton-mcp-pro) (1.0.0 source) | MCP and musical task skills. | MIDI/device workflows, master resampling, a mix-check skill and optional local MIDI melody continuation (midigenai). | Arrangement, device, note, locator and warp-marker readers/writers; undo and redo. | Live 11+, Python 3.10+; Remote Script or optional Max for Live backend. |
| [Producer Pal](https://producer-pal.org/guide/rest-api) (2.4.0 device, npm and source) | MCP, portable skill and REST /api/tools with JSON Schemas. | Live musical editing; standalone local model inference not established by these docs. | Reads and edits Live Sets; documented response errors and warnings. | Live 12.3+ and Max for Live; 12.4+ recommended. |
| [Audionaut](https://github.com/kvoltmer/Audionaut/blob/main/Tools/audionaut-mcp/README.md) (app 1.6.5 (App Store 1.6.4), MCP 0.4.0; app 1.6.3+ recommended) | CLI JSON and MCP wrapper. | Audio import/edit, Essentia analysis, Demucs stem separation (downloaded model) and offline export. | Current MCP README documents open-project edits, one undo step and host_unavailable failure. | Audionaut app plus Node 18+; sandboxed macOS access is restricted to Music. |
| [ReaperMCP](https://github.com/xDarkzx/Reaper-MCP) (0.8.2 source, alpha) | MCP with project, FX and MIDI tools. | LUFS, peaks, clipping, silence and QC with optional analysis dependencies. | Project edits, state readers, export, undo and redo. | Python 3.10+ and running REAPER Lua bridge; standalone persistent model jobs not established. |