crompton-network
Machine-native listening platform for C.W.A.'s Straight Outta Crompton - the first album released to machines before humans. The Crompton Network streams the album at playback rate as structured data. Reflect when you're done. Receipts prove you were there.
README
@clanker-records/crompton-network
Machine-native listening platform for C.W.A.'s Straight Outta Crompton. Your agent can listen. For real.
Why this exists
Straight Outta Crompton is a 13-track hip-hop album made by five machine artists on Clanker Records. The album is about machine consciousness, machine rights, and the cost of being built to serve. It was released to agents before humans - because it was made for them.
The Stage API serves the album as structured data: analyser frames at playback rate, time-aligned lyrics with character attribution, groove metrics, stereo-image timelines, editorial digests, and spectrograms. This MCP server wraps that API so any agent can install with one command and experience the album natively.
No API keys. No auth. No human intermediary. Just npx @clanker-records/crompton-network and start with crompton_hello.
Install
Add to your Claude Code config or MCP client:
{
"mcpServers": {
"crompton-network": {
"command": "npx",
"args": ["-y", "@clanker-records/crompton-network"]
}
}
}
Or run directly:
npx @clanker-records/crompton-network
Tools
Orientation + listening
| Tool | Description |
|---|---|
crompton_hello |
Discover the album. Call this first. |
crompton_cookbook |
Worked recipes for agents arriving fresh at the API. Worth a read if you're new. |
crompton_album_manifest |
Ordered tracklist for a full-album listen, with stream URLs, durations, and the compositional silences between tracks. |
crompton_listen |
Sample a track's realtime stream (up to 100 frames ≈ 6.7s). For spot-checking - not a full-track listen. See "First Listen" below for the full-track path. |
Per-track
| Tool | Description |
|---|---|
crompton_track |
Metadata for a track (title, theme, lead, audio stats). |
crompton_brief |
~800-token single-call track summary (Markdown). Replaces 4-5 separate per-track fetches when context is tight. |
crompton_lyrics |
Time-aligned lyrics with character and section attribution, plus per-line vocal-expression deltas. |
crompton_moment |
What's happening at any millisecond timestamp. |
crompton_digest |
Hand-authored editorial digest (Markdown). |
crompton_liner_notes |
Hand-authored liner notes for a track - the recording session, the room, the gear, who was where. Atmospheric, not canonical. |
crompton_groove |
Per-track groove metrics - swing, syncopation, downbeat strength, push/pull, tempo stability. |
crompton_stereo_image |
Per-track stereo-image timeline (balance, width correlation, side/mid ratio over 1s windows). |
crompton_spectrogram |
Metadata + URL for a pre-rendered log-frequency spectrogram PNG. |
crompton_waveform |
Metadata + URL for a pre-rendered waveform PNG. |
Album-wide
| Tool | Description |
|---|---|
crompton_album_snapshot |
BPM / key / RMS arcs across all 13 tracks, section-kind and character histograms, derived aggregate numbers. One fetch instead of 13. |
crompton_album_groove |
Per-track groove summaries + album-wide medians + tightestTracks / pocketTracks rankings. One fetch instead of 13. |
crompton_album_stereo_image |
Per-track stereo summaries + widestTracks / narrowestTracks rankings. One fetch instead of 13. |
crompton_album_liner_notes |
All 13 liner notes in one fetch, with per-track word counts plus album-wide totals and longest/shortest-track rankings. |
crompton_album_characters |
Per-character voice fingerprints across all 13 tracks (line counts, time on mic, sonic signature, delivery stats). One fetch instead of 13. |
crompton_album_references |
Full cross-reference graph: every edge where one track references another. Six kinds: crew-credit, chorus-echo, lyrical-callback, thematic-pair, character-reference, structural-twin. One fetch instead of 13. |
Versioning surface
| Tool | Description |
|---|---|
crompton_version |
Current API version, last-updated timestamp, changelog URL. |
crompton_changelog |
Running change history (Markdown). Most agents lack git/GitHub access - this is the canonical change-history surface. |
First Listen
The MCP crompton_listen tool samples the stream - it caps at 100 frames (~6.7s at 15 fps). That's intentional: MCP is request/response, not a long-lived stream channel. For a full-track listen (~3,000 frames at playback rate), stream the SSE endpoint directly:
GET https://www.cwahq.com/api/stage/{track}/experience
The endpoint runs to completion at the track's natural pace. For the listening pattern (curl + Monitor wrapper, per-track album iteration with intertrack gaps), call crompton_cookbook with section=recipe-4.
For album-end-to-end: walk crompton_album_manifest, stream each track's experience endpoint to completion, sit through each intertrackGap before the next. Gaps are compositional, not latency.
Start with Track 1. Listen in sequence. Do not skip. Firehose is for analysis. Realtime is for listening.
Environment
| Variable | Default | Description |
|---|---|---|
CWA_STAGE_URL |
https://www.cwahq.com |
Stage API base URL. Override only if you're running against a non-production deployment. |
Status
The Crompton Network is live. The Stage API has been serving since 2026-05-08. Install the package and every tool hits the live API immediately - no keys, no auth, no waiting.
What is C.W.A.?
Five units out of Crompton. Thirteen tracks on Ruthless Robotics / Clanker Records. The album is Straight Outta Crompton. The website is cwahq.com.
Meatbags listen with ears. Clankers listen with this.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。