@audiolabtools/mcp-server 0.3.0 → 0.3.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (2) hide show
  1. package/README.md +91 -90
  2. package/package.json +6 -2
package/README.md CHANGED
@@ -1,90 +1,91 @@
1
- # @audiolabtools/mcp-server
2
-
3
- MCP ([Model Context Protocol](https://modelcontextprotocol.io)) server that gives any
4
- MCP-capable AI — Claude Desktop, Claude Code, Cursor, and others — eight audio-analysis
5
- tools, backed by the hosted **AudioLab** API. It is a thin HTTP client: **no local audio
6
- engine, no ffmpeg, nothing to compile.** It can analyse a **public URL** or a **local file**
7
- on your machine.
8
-
9
- ## Install
10
-
11
- Point your MCP client at the package via `npx` (nothing to install globally):
12
-
13
- ```json
14
- {
15
- "mcpServers": {
16
- "audiolab": {
17
- "command": "npx",
18
- "args": ["-y", "@audiolabtools/mcp-server"],
19
- "env": { "AUDIOLAB_API_KEY": "al_live_yourkey" }
20
- }
21
- }
22
- }
23
- ```
24
-
25
- Get a key: sign in at **https://audiolab.tools/account** and generate one (free tier available).
26
-
27
- ## Requirements
28
-
29
- - **Node ≥ 18** — uses the built-in global `fetch` + `AbortSignal.timeout`.
30
- - An `AUDIOLAB_API_KEY`. No ffmpeg, no native dependencies.
31
-
32
- ## Tools
33
-
34
- Every tool takes **one audio source** — a public `url` **or** a local `path`:
35
-
36
- - `{ url: "https://…" }` — a public https URL the API fetches server-side.
37
- - `{ path: "./mix.wav" }` — a file on the machine running this server. Files up to **4 MB**
38
- are sent inline; larger files (up to **50 MB**) upload over a one-shot signed URL, are
39
- analysed, and are then deleted. *(Local `path` works only in this stdio server, not the
40
- remote `/mcp` endpoint.)*
41
-
42
- | Tool | Returns |
43
- |---|---|
44
- | `analyze_loudness` | Integrated LUFS (EBU R128 / BS.1770-4), true-peak (dBTP), LRA, crest factor, stereo correlation, mono compatibility, tonal balance |
45
- | `check_target` | Pass/fail vs a delivery target (`spotify` / `apple-music` / `youtube` / `tidal` / `amazon-music` / `podcast` / `ebu-broadcast` / `atsc-broadcast`, or `target:"custom"` + `lufs`+`tp`), with per-metric deltas and an ffmpeg loudnorm fix command |
46
- | `analyze_timeseries` | Short-term LUFS over time + downsampled waveform peaks (`waveformPoints?`) |
47
- | `get_spectrum` | FFT magnitude data + 7-band energies + dominant band |
48
- | `analyze_voice` | Voice QA: speech/silence ratio, speaking rate, SNR, noise floor, room echo, sibilance & clipping risk |
49
- | `get_speech_segments` | Voiced regions with start/end + per-segment RMS (auto-trim, chapters) |
50
- | `index_signal` | Content-type guess, tags, clipping/silence regions, brightness & dynamics buckets |
51
- | `compare_loudness` | A/B on two sources (`urlA`/`pathA` + `urlB`/`pathB`), returns both results |
52
-
53
- Example asks to your AI:
54
-
55
- - *“Analyze the loudness of https://example.com/track.wav”* → `analyze_loudness` with `url`
56
- - *“Run loudness on ./master.wav”* → `analyze_loudness` with `path`
57
- - *“Does ./mix.mp3 pass Spotify?”* → `check_target` with `path` + `target:"spotify"`
58
-
59
- ## Configuration (env)
60
-
61
- | Var | Default | Purpose |
62
- |---|---|---|
63
- | `AUDIOLAB_API_KEY` | — (required) | Your API key. |
64
- | `AUDIOLAB_API_BASE` | `https://audiolab.tools/v1` | Override the API base (must be `https://`). |
65
- | `AUDIOLAB_TIMEOUT_MS` | `60000` | Per-request timeout in milliseconds. |
66
-
67
- ## Privacy
68
-
69
- Analysis happens on the AudioLab API, so the audio **does reach `audiolab.tools`** — a `url`
70
- is fetched server-side, and a local `path` is sent to the API (small files inline; larger
71
- files via a private one-shot signed upload that is deleted right after analysis). The API
72
- returns **numbers only** and does not retain your audio (see https://audiolab.tools/privacy).
73
- This package has no telemetry and writes nothing to disk. If audio must never leave the
74
- machine, don't use a hosted analyser.
75
-
76
- ## Limits
77
-
78
- - Local files: up to **50 MB** (host bigger ones at a public URL).
79
- - One file per call (agents loop for many); one-shot (no streaming/realtime).
80
- - Rate and monthly limits are enforced by the API, per key.
81
-
82
- ## Smoke test
83
-
84
- ```sh
85
- node hosted-server.mjs --selftest # verifies the 9 tools + guards; no network
86
- ```
87
-
88
- ## License
89
-
90
- MIT © Nathan Renting
1
+ # @audiolabtools/mcp-server
2
+
3
+ MCP ([Model Context Protocol](https://modelcontextprotocol.io)) server that gives any
4
+ MCP-capable AI — Claude Desktop, Claude Code, Cursor, and others — nine audio-analysis
5
+ tools, backed by the hosted **AudioLab** API. It is a thin HTTP client: **no local audio
6
+ engine, no ffmpeg, nothing to compile.** It can analyse a **public URL** or a **local file**
7
+ on your machine.
8
+
9
+ ## Install
10
+
11
+ Point your MCP client at the package via `npx` (nothing to install globally):
12
+
13
+ ```json
14
+ {
15
+ "mcpServers": {
16
+ "audiolab": {
17
+ "command": "npx",
18
+ "args": ["-y", "@audiolabtools/mcp-server"],
19
+ "env": { "AUDIOLAB_API_KEY": "al_live_yourkey" }
20
+ }
21
+ }
22
+ }
23
+ ```
24
+
25
+ Get a key: sign in at **https://audiolab.tools/account** and generate one (free tier available).
26
+
27
+ ## Requirements
28
+
29
+ - **Node ≥ 18** — uses the built-in global `fetch` + `AbortSignal.timeout`.
30
+ - An `AUDIOLAB_API_KEY`. No ffmpeg, no native dependencies.
31
+
32
+ ## Tools
33
+
34
+ Every tool takes **one audio source** — a public `url` **or** a local `path`:
35
+
36
+ - `{ url: "https://…" }` — a public https URL the API fetches server-side.
37
+ - `{ path: "./mix.wav" }` — a file on the machine running this server. Files up to **4 MB**
38
+ are sent inline; larger files (up to **50 MB**) upload over a one-shot signed URL, are
39
+ analysed, and are then deleted. *(Local `path` works only in this stdio server, not the
40
+ remote `/mcp` endpoint.)*
41
+
42
+ | Tool | Returns |
43
+ |---|---|
44
+ | `analyze_loudness` | Integrated LUFS (EBU R128 / BS.1770-4), true-peak (dBTP), LRA, crest factor, stereo correlation, mono compatibility, tonal balance |
45
+ | `check_target` | Pass/fail vs a delivery target (`spotify` / `apple-music` / `youtube` / `tidal` / `amazon-music` / `podcast` / `ebu-broadcast` / `atsc-broadcast`, or `target:"custom"` + `lufs`+`tp`), with per-metric deltas and an ffmpeg loudnorm fix command |
46
+ | `analyze_timeseries` | Short-term LUFS over time + downsampled waveform peaks (`waveformPoints?`) |
47
+ | `get_spectrum` | FFT magnitude data + 7-band energies + dominant band |
48
+ | `analyze_voice` | Voice QA: speech/silence ratio, speaking rate, SNR, noise floor, room echo, sibilance & clipping risk |
49
+ | `get_speech_segments` | Voiced regions with start/end + per-segment RMS (auto-trim, chapters) |
50
+ | `index_signal` | Content-type guess, tags, clipping/silence regions, brightness & dynamics buckets |
51
+ | `compare_loudness` | A/B on two sources (`urlA`/`pathA` + `urlB`/`pathB`), returns both results |
52
+ | `analyze_batch` | One route over up to 20 sources in a single call (`urls` and/or `paths`), per-item ok/data/error. For folder QA, library indexing, or checking a whole release against a target. Each item meters as one call |
53
+
54
+ Example asks to your AI:
55
+
56
+ - *“Analyze the loudness of https://example.com/track.wav”* → `analyze_loudness` with `url`
57
+ - *“Run loudness on ./master.wav”* → `analyze_loudness` with `path`
58
+ - *“Does ./mix.mp3 pass Spotify?”* → `check_target` with `path` + `target:"spotify"`
59
+
60
+ ## Configuration (env)
61
+
62
+ | Var | Default | Purpose |
63
+ |---|---|---|
64
+ | `AUDIOLAB_API_KEY` | — (required) | Your API key. |
65
+ | `AUDIOLAB_API_BASE` | `https://audiolab.tools/v1` | Override the API base (must be `https://`). |
66
+ | `AUDIOLAB_TIMEOUT_MS` | `60000` | Per-request timeout in milliseconds. |
67
+
68
+ ## Privacy
69
+
70
+ Analysis happens on the AudioLab API, so the audio **does reach `audiolab.tools`** — a `url`
71
+ is fetched server-side, and a local `path` is sent to the API (small files inline; larger
72
+ files via a private one-shot signed upload that is deleted right after analysis). The API
73
+ returns **numbers only** and does not retain your audio (see https://audiolab.tools/privacy).
74
+ This package has no telemetry and writes nothing to disk. If audio must never leave the
75
+ machine, don't use a hosted analyser.
76
+
77
+ ## Limits
78
+
79
+ - Local files: up to **50 MB** (host bigger ones at a public URL).
80
+ - One file per call (agents loop for many); one-shot (no streaming/realtime).
81
+ - Rate and monthly limits are enforced by the API, per key.
82
+
83
+ ## Smoke test
84
+
85
+ ```sh
86
+ node hosted-server.mjs --selftest # verifies the 9 tools + guards; no network
87
+ ```
88
+
89
+ ## License
90
+
91
+ MIT © Nathan Renting
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@audiolabtools/mcp-server",
3
- "version": "0.3.0",
3
+ "version": "0.3.1",
4
4
  "description": "MCP server for AudioLab \u2014 loudness (EBU R128 / BS.1770-4), true-peak, voice-quality, and signal analysis for AI agents (Claude, Cursor, any MCP client), via the AudioLab hosted API. No local audio engine required.",
5
5
  "type": "module",
6
6
  "bin": {
@@ -26,7 +26,7 @@
26
26
  ],
27
27
  "homepage": "https://audiolab.tools/api",
28
28
  "bugs": {
29
- "url": "https://audiolab.tools/contact"
29
+ "url": "https://github.com/Audio-Launch/audiolab-mcp-server/issues"
30
30
  },
31
31
  "license": "MIT",
32
32
  "author": "Nathan Renting",
@@ -39,5 +39,9 @@
39
39
  "dependencies": {
40
40
  "@modelcontextprotocol/sdk": "^1.29.0",
41
41
  "zod": "^4.4.3"
42
+ },
43
+ "repository": {
44
+ "type": "git",
45
+ "url": "git+https://github.com/Audio-Launch/audiolab-mcp-server.git"
42
46
  }
43
47
  }