Figma Context MCP
Give your coding agent access to your Figma data. Implement designs in any framework in one-shot.
作者提供的说明(英文原文)
Unofficial. Status, AMS filament, print control and maintenance for Bambu Lab 3D printers via LAN
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
BAMBU_IP, BAMBU_SERIAL, BAMBU_ACCESS_CODE, BAMBU_READ_ONLY标为机密的值是你账户的钥匙。千万不要分享给别人,也不要粘贴到聊天里。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json
{
"mcpServers": {
"for-bambu": {
"command": "npx",
"args": [
"-y",
"mcp-server-for-bambu"
],
"env": {
"BAMBU_IP": "your-value",
"BAMBU_SERIAL": "your-value",
"BAMBU_ACCESS_CODE": "your-value",
"BAMBU_READ_ONLY": "your-value"
}
}
}
}在终端中运行这条命令
claude mcp add for-bambu --env BAMBU_IP=your-value --env BAMBU_SERIAL=your-value --env BAMBU_ACCESS_CODE=your-value --env BAMBU_READ_ONLY=your-value -- npx -y mcp-server-for-bambu运行前请先填写空着的值。
粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"for-bambu": {
"command": "npx",
"args": [
"-y",
"mcp-server-for-bambu"
],
"env": {
"BAMBU_IP": "your-value",
"BAMBU_SERIAL": "your-value",
"BAMBU_ACCESS_CODE": "your-value",
"BAMBU_READ_ONLY": "your-value"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"for-bambu": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"mcp-server-for-bambu"
],
"env": {
"BAMBU_IP": "your-value",
"BAMBU_SERIAL": "your-value",
"BAMBU_ACCESS_CODE": "your-value",
"BAMBU_READ_ONLY": "your-value"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
这个服务器在你的电脑上运行,而 ChatGPT 只能连接互联网上的服务器。请选择其他应用。
你的选择通过 Cookie 保存在这台设备上,无需账户。
BAMBU_IP envBAMBU_SERIAL envBAMBU_READ_ONLY env这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
它在你的电脑上运行;安装后,工具列表会显示在你的 AI 应用中。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月7日
注册名称已通过 GitHub 账户 b84330808 验证,该账户拥有此仓库
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
它在你的电脑上运行,没有互联网端点
在你的电脑上运行,改用本地设置
它在你的电脑上运行;我们只能读取互联网端点的工具列表
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
你需要填写:BAMBU_ACCESS_CODE
它在你的电脑上运行。你的 AI 应用会在需要时启动它,数据不会经过作者的服务器(除非它自己调用在线服务)。
GitHub 仓库最后更新于–,最新版本(0.3.0)于12天前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。你的电脑上还需要有 Node.js。
同一分类中有:Figma Context MCP, Claude Real Video, Figma MCP Server
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Give your coding agent access to your Figma data. Implement designs in any framework in one-shot.
Let any LLM watch a video locally — and search everything it has ever watched.
The Figma MCP server brings Figma design context directly into your AI workflow.
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
Ask Claude (or Cursor, or VS Code) what your Bambu Lab printer is doing, and have it pause, preheat or log maintenance, over your LAN with no cloud.
An MCP server for Bambu Lab 3D printers. It talks to the printer directly over its local MQTT interface. Built and tested against the P1S; other MQTT-speaking models (P1P, X1C, A1) should work.
⚠️ Unofficial — not affiliated with Bambu Lab
This is a community-built project. It is not affiliated with, endorsed by, or sponsored by Bambu Lab / Shenzhen Tuozhu Technology Co., Ltd. "Bambu Lab", "Bambu Studio", "P1S", "P1P", "X1C", and "AMS" are trademarks of their respective owners; references in this project are descriptive only (nominative fair use), made for compatibility purposes. Provided "as is" with no warranty.
*Real tool calls and output, run against the built-in printer simulator (npm run sim). Regenerate with scripts/demo/; docs/demo-script.md covers recording a session with a real printer.*
| Live status | State, current stage, progress, layer, time left, nozzle/bed/chamber temperatures, Wi-Fi signal |
| AMS | Filament type, brand, color and remaining % per slot, plus the active slot |
| Health | Active HMS warnings decoded to module and severity, each with its Bambu wiki link |
| Control | Pause, resume, cancel, chamber light, speed profile, preheat, AMS slot switch, raw G-code |
| Maintenance ledger | Counts print-hours locally and tells you when rails, carbon rods or the hotend are due |
| Guided setup | No config editing: tell Claude your IP, serial and access code; it tests them before saving |
| Safety | Cancel, G-code and AMS switch need explicit confirmation; read-only mode; LAN addresses only |
| Works anywhere MCP does | Claude Code plugin, Claude Desktop, Cursor, VS Code, any stdio MCP client |
| Printer simulator | npm run sim starts a fake printer for trying it out, demos and tests |
You need Node.js 20+ and a printer on the same network with LAN Mode Liveview turned on.
Install the plugin (MCP server plus a short skill that teaches Claude the confirmation rules):
/plugin marketplace add b84330808/mcp-server-for-bambu
/plugin install bambu-printer@mcp-server-for-bambu
Or add just the MCP server:
claude mcp add --transport stdio --scope user bambu -- npx -y mcp-server-for-bambu
Open Settings → Developer → Edit Config (%APPDATA%\Claude\claude_desktop_config.json on Windows, ~/Library/Application Support/Claude/claude_desktop_config.json on macOS) and add:
{
"mcpServers": { "bambu": { "command": "npx", "args": ["-y", "mcp-server-for-bambu"] } } }
Fully quit Claude Desktop (tray icon → Quit) and reopen it.
Or put this in ~/.cursor/mcp.json (all projects) or .cursor/mcp.json (one project):
{
"mcpServers": { "bambu": { "command": "npx", "args": ["-y", "mcp-server-for-bambu"] } } }
code --add-mcp "{\"name\":\"bambu\",\"type\":\"stdio\",\"command\":\"npx\",\"args\":[\"-y\",\"mcp-server-for-bambu\"]}"
Or add it to .vscode/mcp.json. This version passes the credentials as env vars and asks for the access code as a hidden input:
{
"inputs": [ { "type": "promptString", "id": "bambu-access-code", "description": "Bambu LAN access code (Settings > WLAN > LAN Mode Liveview)", "password": true } ], "servers": { "bambu": { "type": "stdio", "command": "npx", "args": ["-y", "mcp-server-for-bambu"], "env": { "BAMBU_IP": "192.168.1.50", "BAMBU_SERIAL": "01P00A1234567890", "BAMBU_ACCESS_CODE": "${input:bambu-access-code}" } } } }
All of these are in examples/.
On first start the server is in setup mode and offers one tool, setup_printer. Read three values off the printer's touchscreen:
Settings → WLAN → tap the connected networkSettings → Device → Device Info (P1S: 01P00A…)Settings → WLAN → LAN Mode Liveview, 8 characters, not the Wi-Fi password. LAN Mode Liveview must be ON.Then tell your assistant:
"Set up my Bambu printer. IP is 192.168.1.50, serial 01P00A1234567890, access code 12345678."
setup_printer checks the format, connects to the printer, and saves the credentials only if the connection works. Restart the server once (Claude Desktop: quit and reopen; Claude Code: /mcp → reconnect bambu) and the printer tools appear.
You can skip this and pass BAMBU_IP, BAMBU_SERIAL and BAMBU_ACCESS_CODE as env vars instead, as in the VS Code example.
| Tool | What it does | Asks first? |
|---|---|---|
get_status | State, stage, temperatures, progress, layer, HMS warnings, Wi-Fi | |
get_ams_info | Per-slot filament type / color / remaining %, active slot | |
list_active_warnings | Active HMS warnings with links to Bambu's wiki | |
get_maintenance_status | Print-hours and lube / clean / inspect warnings | |
forecast_maintenance | When each maintenance task will be due at your recent print rate | |
get_recent_prints | Last N print sessions: name, duration, finish time | |
mark_maintenance_done | Log a maintenance task (optionally back-dated by hours_ago) | |
pause_print / resume_print | Pause and resume the current print | |
set_light | Chamber light on/off | |
set_speed | silent / standard / sport / ludicrous | |
preheat | Set nozzle and/or bed target temperature | |
stop_print | Cancel the print. Not reversible | yes (confirm: true) |
run_gcode | Send raw G-code (homing, Z moves, manual unload…) | yes (confirm: true) |
change_active_tray | Switch the AMS to another slot | yes (confirm: true) |
Every tool also carries MCP annotations (readOnlyHint, destructiveHint), so clients that ask before running tools can treat the read tools and the destructive ones differently.
This server can move a heated toolhead and cancel hours of printing. It is built to make that hard to do by accident:
stop_print, run_gcode and change_active_tray refuse to run unless called with confirm: true, and their descriptions tell the model to set it only after you agree in plain words. Keep your client's own tool-approval prompts on as a second check.BAMBU_READ_ONLY=1 and the server registers only the read tools. Nothing can pause, cancel, heat or move the printer. A good default for a shared machine or a demo.setup_printer rejects addresses outside private ranges (192.168.x.x, 10.x.x.x, 172.16–31.x.x, plus link-local, loopback and the 100.64/10 range Tailscale uses). Never forward the printer's MQTT port to the internet.preheat and M104/M140 leave the heaters at the target until something changes them. Turn them off (set 0) before you walk away.| Env var | Default | Meaning |
|---|---|---|
BAMBU_IP, BAMBU_SERIAL, BAMBU_ACCESS_CODE | from the saved .env | Printer credentials. Env vars win over the saved file |
BAMBU_READ_ONLY | 0 | 1 registers only the read tools |
BAMBU_CONFIG_DIR | see below | Where the .env and the maintenance ledger live |
MAINTENANCE_LEDGER_PATH | /maintenance.json | Ledger file location |
The config directory is:
%APPDATA%\mcp-server-for-bambu\~/Library/Application Support/mcp-server-for-bambu/~/.config/mcp-server-for-bambu/ (or $XDG_CONFIG_HOME)Versions up to 0.2.0 kept .env and maintenance.json inside the installed package, where npx updates could delete them. 0.3.0 still reads an old .env from there and copies an old ledger into the config directory on first start.
The ledger persists to a local JSON file and survives crashes (active sessions are checkpointed every 5 elapsed minutes). Thresholds are in src/types.ts:
| Task | Interval |
|---|---|
| Clean carbon rods | 100 print-hours |
| Lubricate linear rails | 200 print-hours |
| Inspect hotend / nozzle wear | 500 print-hours |
A task flips to DUE_SOON (🔔) within 10 hours of the threshold and OVERDUE (⚠️) past it.
npm run sim (from a clone, after npm install) starts a simulated P1S halfway through a Benchy, with a 4-slot AMS, on mqtts://127.0.0.1:8883. It answers status requests and applies pause, resume, cancel, speed, preheat, light and AMS commands. Point the server at it:
BAMBU_IP=127.0.0.1 BAMBU_SERIAL=01P00ASIMULATOR BAMBU_ACCESS_CODE=12345678 node dist/index.js
Set SIM_PORT to use another port, then pass the same value as BAMBU_MQTT_PORT. The test suite uses the same simulator.
git clone https://github.com/b84330808/mcp-server-for-bambu.git
cd mcp-server-for-bambu npm install npm test # builds, then runs unit + end-to-end tests against the simulator npm run dev # tsc --watch npm run sim # simulated printer
To run your local build from an MCP client, point it at node /absolute/path/to/mcp-server-for-bambu/dist/index.js.
src/
index.ts MCP server bootstrap + tool definitions config.ts Env / .env loading, LAN address check paths.ts Per-user config directory mqtt-client.ts TLS MQTT client for the printer (port 8883, self-signed cert) setup.ts setup_printer: validate, test the connection, save maintenance.ts Print-time ledger with crash-safe atomic writes + recovery hms.ts Bambu HMS error code → human label stages.ts Stage / speed / tray code → human label commands.ts MQTT command payloads sim/printer.mjs Simulated printer (aedes broker over TLS) plugins/ Claude Code plugin; .claude-plugin/marketplace.json lists it server.json MCP Registry metadata
Protocol notes:
device/{serial}/reportdevice/{serial}/request with {"pushing":{"sequence_id":"...","command":"pushall"}}bblp, password = access code, a random client只显示了开头部分;其余内容请在仓库中阅读