PowerTree MCP Server
Enables AI assistants to perform electronic circuit power tree analysis, including opening projects, solving margins, validating designs, editing elements, managing waivers, and exporting reports.
README
PowerTree
Electronic circuit power tree analysis — a fully offline desktop app for budgeting power from source to load with min/typ/max corners, margin analysis, a live flowchart, and report-grade exports (PDF / HTML / Excel / CSV / PNG).
— PySide6 · Windows-first · no accounts, no internet.
pip install powertree
powertree-gui :: desktop app (opens with a Zynq demo board)
powertree --help :: full CLI
What it does
- Projects hold many power trees (
.ptproj, versioned JSON, self-contained). - Elements: one Source per tree (V min/typ/max + current or power limit), Converters (efficiency %, Vout corners, quiescent current, optional output limit, pass-through output rail for sub-loads), Loads (current- or power-type, typ/peak value, allowed input-voltage window for margin checks), Series elements (resistance, bounded so the math never breaks). Every element carries the same metadata: name, signal name, ref des, part number, pin(s), datasheet link, notes.
- Blocks group elements visually (e.g. one IC with Icc + Iq loads) and show aggregate power.
- Bottom-up solver, auto-refresh: every edit re-solves all three corners (damped fixed-point for series-R × power-load interaction) and repaints instantly.
- Margin analysis: source/converter limit usage, load under/over-voltage vs its allowed window, step-down/boost sanity, collapsed-rail detection — surfaced as findings, node badges, list statuses, and report sections.
- Views: color-coded flowchart (90° routed arrows with rail labels, legend, collapse/expand chips, no overlaps, power text on every card) + hierarchical list view. Layouts: top-down, left-right, or custom drag-to-place.
- Search (Ctrl+F) across name/signal/refdes/part/pins/notes — highlights the canvas and filters the list.
- Documentation notes: hierarchical markdown notes with embedded images, linked to elements — capture where every number came from.
- Exports: PDF report (flowcharts + tables + margins + full notes appendix),
macro-enabled Excel report (
.xlsmwith outline collapse + VBA navigation macros; falls back to.xlsx+ importable.baswhen VBA trust is off), HD PNG flowchart, notes → Markdown / HTML / PDF.
Install & run
From PyPI (recommended):
pip install powertree :: add powertree[mcp] for the AI/MCP server
powertree-gui :: GUI
powertree info my.ptproj :: CLI (same entry: powertree --help)
From source:
git clone https://github.com/bin2ai/powertree
cd powertree
py -3.12 -m venv .venv
.venv\Scripts\pip install -r requirements.txt
.venv\Scripts\python.exe main.py :: or double-click PowerTree.bat
Standalone Windows installer (no Python needed) — build
dist\PowerTree-Setup.zip with installer\build_installer.ps1, unzip, run
install.bat (per-user install, Start-Menu/Desktop shortcuts, uninstaller;
installer\PowerTree.iss builds a setup.exe when Inno Setup is present).
Onboarding: press F1 in the app (Quick start), or read
docs/QUICKSTART.md and
docs/USER_GUIDE.md. A full feature tour with
screenshots: artifacts/PowerTree_Feature_Overview.pdf.
The app opens with a built-in demo (a realistic Zynq-7000 carrier board) so every feature is visible immediately.
Four ways to use it
| Mode | How |
|---|---|
| GUI | powertree-gui — flowchart, list, properties, states, notes, search, waivers |
| CLI | powertree info|solve|validate|nets|headroom|growth|bom|search|export|templates|demo … — JSON with --json; validate exits non-zero on violations (--strict fails on warnings too) |
| Excel | macro-enabled report export, native outline collapse, live formulas, States/Parts sheets |
| AI / MCP | powertree-mcp — 16 tools (open/solve/validate/edit/waive/export); see examples/mcp.json.example for Claude Code / Desktop registration |
powertree info examples\DemoBoard.ptproj
powertree validate examples\DemoBoard.ptproj --strict && echo margins clean
powertree growth examples\DemoBoard.ptproj :: +N% load headroom
powertree export bundle examples\DemoBoard.ptproj -o out\ :: all formats
Tests
.venv\Scripts\python.exe -m pytest tests -q
14 tests cover the solver math (hand-checked corner cases incl. the series-R + power-load fixed point), model constraints, and file-format round-tripping.
Layout
main.py launcher
src/powertree/
model/ elements, bottom-up solver (calc.py), .ptproj serialization
ui/ main window, flowchart canvas, tidy-tree layout, list view,
properties, notes, dark theme
export/ PDF report, Excel (.xlsm/COM + .xlsx), HD PNG, notes MD/HTML/PDF
sampledata.py demo project
tests/ pytest suite
examples/ DemoBoard.ptproj
notes/ Obsidian knowledge base (project docs, not app data)
Excel macros note
.xlsm export embeds VBA (expand/collapse outline, next-finding navigation) through
Excel COM, which requires File → Options → Trust Center → Trust Center Settings →
Macro Settings → “Trust access to the VBA project object model”. Without it you get
an identical .xlsx plus PowerTree_Macros.bas to import manually (Alt+F11 →
File → Import).
Roadmap ideas
Resistive loads · operating modes / duty-cycle scenarios · undo-redo · temperature derating · CSV/BOM import · converter component library · rail sequencing checks · per-corner load values.
Scaffolded 2026-08-06. See CLAUDE.md for how Claude works in this project.
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。