One Interface. Multiple Engineering Capabilities.
The goal is one interface through which your coding agent can discover and coordinate AIPE’s independent engineering agents, databases and tools.
Paste this into Codex, Claude Code or another coding agent. It points the agent to aipe.md, AIPE’s public, agent-readable index of engineering capabilities.
Read https://aipel.co.uk/aipe.md, identify the AIPE engineering capabilities relevant to my task, and help me use the appropriate tools and agents.
Capability discovery Available
Your agent reads the index, finds the AIPE capabilities that fit your task and follows each one’s own documentation. Each tool is still set up on its own, as its repository describes.
Installable plugin and orchestration Planned
One installation that sets up AIPE agents and coordinates them across a task. In design: no plugin package has been released.
Each agent lives in its own repository. Copy prompt gives your coding agent an instruction to use that agent; it does not install anything. Where an agent is also a native Codex and Claude Code plugin, its card links to the installation steps: install the plugin first, then paste the prompt.
Turn circuit descriptions into clear, electrically checked SVG schematics, with a JSON validation report.
Install the Codex or Claude Code plugin first, then paste the prompt. Install steps →
Using AIPE Sketcher, generate a Dual Active Bridge converter schematic, validate the electrical connections, and export the result as SVG.
AI-assisted power-electronics PCB design in KiCad, with native ERC/DRC checks and a recorded engineering review.
Read https://github.com/FulongLi/AIPE-PCB-Agent and follow its README and docs/agent-workflow.md to help me with my power-electronics PCB task. Inspect my environment first and ask before installing any software. Use KiCad's native ERC and DRC checks, state every assumption, and list everything that still needs human engineering review.
Turn a research idea into an evidence-linked paper plan and a LaTeX workspace built on official IEEE templates.
Read https://github.com/FulongLi/AIPE-IEEE-Paper-Agent and follow its AGENTS.md and core/workflows/idea-to-paper.md. My research idea is: <describe your idea>. Develop a paper plan with testable questions and evidence gaps, then set up a manuscript workspace for my target IEEE venue. Never invent results or references; mark missing evidence as TODO.
AIPE Academy ships as a Codex skill and a Claude Code plugin that teach from the same open curriculum: four domains, prerequisite-aware learning paths, reproducible Python labs, MIT OpenCourseWare scope references and a learning record that stays on your machine. Beyond your coding agent it needs only Python 3.10; no API subscription, proprietary software or cloud database. Available
/plugin install aipe-academy --marketplace FulongLi/AIPE-Academy
Then run /aipe-academy:learn, or just ask: “Teach me electrical engineering from the beginning.”
git clone https://github.com/FulongLi/AIPE-Academy
Open the folder in Codex; the aipe-academy skill is discovered from .agents/skills/. Type $aipe-academy or say “Guide me through AIPE Academy from the beginning.”
Step-by-step instructions, what has been tested on each platform and known limitations: docs/plugin-installation.md ↗. Every lesson on this site links to its canonical source in the same repository.
Any agent that can read a public web page and work in your project, including Codex and Claude Code. Some agents may ask for permission to access GitHub or run local commands.
No. Reading AIPE’s public pages and repositories does not send project files to AIPE Labs. Information entered into a third-party agent is handled under that provider’s terms and your organisation’s policies.
No. Every AIPE capability states its limitations. Use it to speed up discovery and analysis, then verify important results against original datasheets, models, standards, simulations and measurements.