Skip to content

Adam Engelbrecht (GitHub: RFingAdam) — which repos matter to MagicON (2026-10-06)

Sources: GitHub API repo list (gh api users/RFingAdam/repos, read 2026-10-06) and each repo's own README. Nothing was installed or called. Star counts are from the API on that date.

Useful

Repo What it is Why it matters to us
lineforge (AGPL-3.0 per GitHub; its README text also says "GPLv3", 1 star) Transmission-line calculator over MCP: microstrip, stripline, CPWG, diff pairs, any 2D cross-section; claims ±2 % vs IPC-2141 and openEMS. README says 14 tools, the server code defines 21 (incl. target_z0, laminate_lookup, rf_path_budget = pad/trace return loss). "Pre-alpha: not yet on PyPI" A competitor we had missed — it overlaps our impedance solving. Now added to mcp-competitor-comparison-2026-10-06.md (ranked #10 of 11). It does not design a stackup: it reduces the layers around one trace to one equivalent dielectric (series_reduce, read in its code). Picking layer count, materials and thicknesses for a whole board stays ours
drawio-engineering-mcp (AGPL-3.0, 4 stars) draw.io MCP, 10 tools: create_rf_block_diagram (from JSON blocks, with its own gain/NF cascade), create_pcb_stackup (layers + materials), EMC test setups A partner, not a rival. It draws; MagicON designs. An AI assistant could pass a MagicON chain or stackup result into it to get a diagram. Small overlap: it does its own cascade math
Adam himself RF/antenna engineer who builds and uses MCP servers ("Automating workflows with AI") Exactly the "MCP / AI-assistant user" our plan says ties go to. A possible early user or feedback source. Contacting him is the owner's call

Already covered

  • mcp-ltspice-qucs — already ranked #2 in our comparison doc.
  • mcp-pcb-emcopilot — already ranked #5. His profile now says the PCB EMC review tooling "stay private", which fits the 404. The eng-mcp-suite link our comparison doc used as a source also returns 404 now (gh api repos/RFingAdam/eng-mcp-suite/readme, 2026-10-06); the doc now links the Glama listing instead.

Small or no use

Repo Why not
mcp-openems Antenna design (patch, dipole, horn, helix) by formula + openEMS script output. Would only overlap our design_patch_antenna, which is not on MCP yet (#176)
rf-interference-calculator Streamlit web app for multi-radio IMD/harmonics; not an MCP server; coexistence is not our scope
mcp-nec2-antenna HF/VHF/UHF wire antennas (Yagi, dipole) — we don't design antennas like these
RFlect, gnss-toolkit Analyse MEASUREMENTS of finished hardware (chamber, VNA, GNSS fixes); we design before hardware exists
mcp-blender, mcp-remote-access, vegas-director-mcp, tactical-deck, agentarchy, BMO, hermes-agent, Fabric, ESP32-DIV, VPN-AP, helium-weather-station, claude-skill-homeassistant Not RF design (3D modelling, SSH, video editing, games, Linux distro, forks, hobby projects)

Licence note

Every RF repo of his is AGPL-3.0. We must not copy their code into MagicON (AGPL would force our server code open). Running one as a separate tool, or linking to it, is normally fine. This is a plain reading of the licence, not legal advice.