I Let AI Route My PCB: 24-Sensor Line Follower in KiCad (KiCad MCP + Freerouting)

I let AI route my PCB! In this video I design two custom PCBs for a competition line follower robot in KiCad: a 24-sensor QRE1113 IR sensor array and a main board with an Arduino Nano, OLED display, motor driver and buzzer. I drew the schematic myself, but the routing is done by Claude Code using the KiCad MCP server and Freerouting. Then we verify everything with ERC, DRC and a Gerber check before sending the boards to JLCPCB. 🤖⚡

If you've ever wondered whether AI can actually help with PCB design, this is a real project, not a toy example. You'll see where AI saves hours (routing, copper zones, track widths) and where you still need to do the work yourself (schematic, component choice, footprints and the final checks).

New to KiCad? Start with my free KiCad course (5 episodes, beginner friendly):
📚 https://www.youtube.com/playlist?list=PLCMNOSPfM9J1dJcvVuaiutZCnwWLF5l_3
➡️ Generating Gerbers & ordering from JLCPCB: https://www.youtube.com/watch?v=lUr3CgoCkj0

⏱️ Timestamps
0:00 Intro: line follower PCBs + AI routing
0:38 Tour: sensor array and main board (3D view)
1:27 Why two separate PCBs?
2:02 Why 24 sensors? QRE1113 vs TCRT5000
2:49 Using two multiplexers for 24 channels
3:41 Net labels for a clean schematic
4:06 How the QRE1113 IR sensor works (330Ω + 10kΩ)
5:51 Component sourcing & footprints
6:32 Branding & M2 mounting holes
6:44 Auto-routing with Freerouting
7:01 KiCad MCP server setup (Node.js, Python, Freerouting)
7:46 Letting Claude Code route the PCB in VS Code
8:12 Don't fully trust AI: do the schematic & ERC yourself
8:37 Decoupling capacitors
8:56 Main board: Arduino Nano, OLED & motor driver (SnapEDA)
9:41 Voltage regulator & why I added a fuse
10:10 Buttons for PID tuning + buzzer
10:40 Connectors for the sensor array, power & motors
11:17 AI-routed main board, copper zones & track widths
12:06 DRC, ERC & Gerber check
12:46 Sent to JLCPCB + what's next

🧩 What's on the boards
IR sensor array PCB:
• 24× QRE1113 reflective IR sensors
• 24× 330Ω resistors (IR LED current limit)
• 24× 10kΩ resistors (phototransistor pull-up)
• 2× 16-channel analog multiplexer modules (sensors 1–12 and 13–24)
• 100nF decoupling capacitor, 8-pin connector, M2 mounting holes

Main board PCB:
• Arduino Nano
• OLED display
• Motor driver module
• Voltage regulator
• Fuse for over-voltage protection
• 3 buttons (up / down / OK) for PID tuning menus
• Buzzer, pin headers, screw terminals for power, motors and the main switch

⚙️ My AI PCB routing workflow
1. Draw the full schematic yourself and use net labels to keep it clean
2. Assign footprints (check the parts are actually available in your country!)
3. Run ERC and fix the real errors
4. Place the parts in the PCB editor
5. Let Claude Code + the KiCad MCP server + Freerouting do the routing
6. Review track widths, copper fill zones and the ground plane
7. Run DRC, then check the Gerber files before ordering

AI is great at the slow, repetitive part. The schematic, the component choice and the ERC/DRC checks are still your job.

❓ FAQ
Why two PCBs instead of one? The sensor array can be reused on any other line follower robot, and if I make a mistake I only remake one small board, not everything.
Why 24 sensors? QRE1113 sensors are small, so you need more of them than TCRT5000 (usually 8 to 16). 16-sensor arrays are already easy to buy here, so I went with 24.
Why a fuse? The motor driver is happy around 12V. If the supply goes higher, the fuse cuts power to protect the Arduino, OLED, regulator and driver.
Do I need Claude Code? Any AI agent that supports MCP should work. You also need Node.js, Python, KiCad and Freerouting installed.

🔗 Tools & references
KiCad MCP server: https://github.com/mixelpixx/KiCAD-MCP-Server

Freerouting: https://github.com/freerouting/freerouting/releases

IR sensor reference: https://www.pcbway.com/project/shareproject/Solunar_is_a_16_array_IR_sensor_designed_for_line_following_robots_It_has_high_1f3d7406.html

SnapEDA (footprints & 3D models): https://www.snapeda.com
KiCad: https://www.kicad.org

🎥 More builds from the channel
Obstacle Avoidance Car With Arduino, Full DIY Guide: https://www.youtube.com/watch?v=d2eHqq3yDFI

Easiest Arduino RC Car Build for Under $50: https://www.youtube.com/watch?v=kY1y4Mbrsyw

I Turned My ESP32 into an AI Assistant (No Coding!): https://www.youtube.com/watch?v=K5lOT7_UyMA

Next video: soldering everything and building the complete PID line follower robot. Subscribe so you don't miss it! 🔔
Any questions about the design? Drop them in the comments 👇

👋 About Roborear
I build robots, ESP32 and Arduino projects, and custom PCBs, and I share every step so you can build them too. New videos cover KiCad, sensors, microcontrollers and full robot builds from schematic to soldering.

#kicad #pcbdesign #linefollower #arduino #robotics #ai