seismo-mcp
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@seismo-mcpBandpass filter 1-5 Hz on waveform.sac and plot it."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
seismo-mcp
Let AI assistants process seismic waveforms by chatting — no code, no scripts.
让 AI 助手通过对话处理地震波形——不写代码,不写脚本。
What is this?
seismo-mcp is a bridge between AI assistants (like Claude, Cursor) and the software seismologists actually use: ObsPy, CWP/SU, and SAC.
Normally, to filter a seismogram you open a terminal and type something like
sac commands or write a Python script with ObsPy. With seismo-mcp
installed, you just tell the AI what you want in plain language, and it
calls the right tool for you.
A 30-second example
Install it once. Then, in Claude Desktop:
You: Here's
event.sac. Bandpass filter 0.5–2 Hz, remove the mean, then cut −5 to 30 seconds and show me the first trace.
The assistant calls the sac-mcp tools behind the scenes — it reads the
header, filters, preprocesses, cuts, and plots — and hands you back the
processed file and a plot. You never wrote a line of code.
you: "bandpass 0.5-2 Hz, demean, cut -5 to 30 s, plot"
│
▼ (the assistant decides which tools to call)
┌─────────────────────────────────────────┐
│ seismo-mcp servers (already installed) │
│ sac-mcp → sac_filter / preprocess / │
│ cut / plot │
└─────────────────────────────────────────┘
│
▼ (the server runs the real software)
SAC / ObsPy / CWP-SU (your local install)
│
▼
processed file + plot, handed back to youRelated MCP server: IMAS Codex
Why would I want this?
For researchers & students — skip the "how do I write this command again?" friction. Describe the processing, get the result.
For multi-toolchain workflows — ObsPy, SU, and SAC speak different dialects; seismo-mcp gives the assistant one consistent way to reach all three.
For teaching & demos — show seismology processing live, in plain language, without a screen full of shell commands.
How do I use it?
1. Install one or more servers
Pick only what you need. Each server is independent.
Server | Use it if you work with | Install command |
obspy-mcp | ObsPy / general waveform I/O |
|
cwp-su-mcp | CWP/SU reflection processing |
|
sac-mcp | SAC waveform analysis |
|
gmt-mcp | Maps & plots (PyGMT) |
|
Prerequisite: the relevant software must be installed on your computer (ObsPy is auto-installed by
obspy-mcp; CWP-SU, SAC, and the GMT binary must be installed separately). See the install guide for details.Note:
uvxcomes from uv — install it withbrew install uv(macOS/Linux) if you don't have it. The packages will be on PyPI soon; until then see INSTALL.md for the from-source method.
2. Restart your AI assistant and just ask
After installing, restart Claude Desktop (or your MCP-compatible client). Then describe what you want — the assistant will use the tools automatically.
Example things to say:
"Read
shot.su, how many traces and what's the sample interval?""Bandpass this SAC file 2–10 Hz and plot it."
"Convert
trace.sacto MiniSEED.""Apply AGC with a 0.5 s window to
gather.su, then sort by offset."
What's inside?
Three independent servers, each wrapping one toolchain:
Server | What it can do |
Read any waveform format (SAC/MSEED/SEG-Y…), filter, preprocess, resample, convert formats, plot. | |
Seismic Unix trace ops: read headers, filter, gain (AGC), window, sort, set headers. | |
SAC waveform ops: list headers, preprocess, filter, cut, merge, remove instrument response. | |
Maps & plots: basemaps, coastlines, epicenter/station maps, x-y plots, labels. |
Who is this for?
Seismologists who want a faster way to drive their existing tools.
Students learning seismology — see what each operation does, in plain words.
Anyone curious about connecting scientific software to AI assistants.
Learn more
📖 Install guide — step-by-step setup (ObsPy, CWP-SU, SAC, uv).
🛠 Design & technical details — why one server per toolchain, the three wrapping patterns, what gotchas were solved.
🔧 Each server's README — full tool tables and examples.
Status
Working and tested against real local installs of ObsPy / CWP-SU / SAC / PyGMT. Four servers, ~28 tools. Feedback and contributions welcome.
License
MIT — see LICENSE.
中文
这是什么?
seismo-mcp 是一座桥,把 AI 助手(Claude、Cursor 等)和地震学家真正在用的 软件连起来:ObsPy、CWP/SU、 SAC。
平时,要处理一个地震波形,你得打开终端敲 sac 命令,或用 ObsPy 写 Python 脚
本。装上 seismo-mcp 后,你只需用大白话告诉 AI 你要什么,它就会替你调用相应
的工具。
一个 30 秒的例子
安装一次。然后在 Claude Desktop 里:
你: 这是
event.sac。带通滤波 0.5–2 Hz,去均值,截取 −5 到 30 秒, 然后画第一道给我看。
助手会在后台调用 sac-mcp 的工具——读头、滤波、预处理、截取、绘图——然后把
处理好的文件和图交给你。你全程没写一行代码。
我为什么要用?
研究者和学生——跳过"这个命令怎么写来着"的折腾,描述你要的处理,直接拿 结果。
跨工具链的工作流——ObsPy、SU、SAC 各说各的方言;seismo-mcp 给了助手一 个统一的方式触达三者。
教学和演示——用大白话现场展示地震处理过程,不用满屏的 shell 命令。
怎么用?
1. 安装你需要的 server(按需,互相独立)
Server | 适用场景 | 安装命令 |
obspy-mcp | ObsPy / 通用波形读写 |
|
cwp-su-mcp | CWP/SU 反射地震处理 |
|
sac-mcp | SAC 波形分析 |
|
gmt-mcp | 地图与绘图(PyGMT) |
|
前提: 相应软件需已装在本机(ObsPy 会由
obspy-mcp自动安装;CWP-SU、 SAC 和 GMT 二进制需自行安装)。详见安装指南。
2. 重启 AI 助手,直接开口
装好后重启 Claude Desktop(或其他支持 MCP 的客户端),然后用大白话描述需求, 助手会自动调用工具。
可以这样说:
"读一下
shot.su,多少道?采样率多少?""把这个 SAC 文件带通滤波 2–10 Hz,然后画出来。"
"把
trace.sac转成 MiniSEED。""对
gather.su做 0.5 秒窗口的 AGC,然后按 offset 排序。"
里面有什么?
三个独立 server,各封装一个工具链:
Server | 能做什么 |
读各种波形格式(SAC/MSEED/SEG-Y…)、滤波、预处理、重采样、格式转换、绘图。 | |
地震 Unix 道处理:读头、滤波、增益(AGC)、时窗、排序、设头。 | |
SAC 波形操作:读头、预处理、滤波、截取、合并、去仪器响应。 | |
地图与绘图:底图、海岸线、震中/台站分布、x-y 图、文字标注。 |
想了解更多
📖 安装指南 —— 分步设置(ObsPy、CWP-SU、SAC、uv)。
🛠 设计与技术细节 —— 为什么一工具链一 server、三种封装模式、解决了哪些坑。
🔧 各 server 的 README —— 完整工具表和示例。
状态
已在真实安装的 ObsPy / CWP-SU / SAC / PyGMT 上测试通过。四个 server、约 28 个 工具。欢迎反馈和贡献。
许可证
MIT —— 见 LICENSE。
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/AnnanRen/seismo-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server