STEP 01

Connect the hub

Pair a SPIKE Prime hub over Bluetooth. The app introspects what is attached: motor types, sensor models, battery level, firmware version, and signal strength.

  • Bluetooth Low Energy pairing in seconds
  • Auto-detect motors, color, distance, and force sensors
  • Surface battery, firmware, and signal as live state
  • Disconnect / reconnect tolerated mid-skill
BLUETOOTH SCAN
SPIKE Prime · #A47F
FW 1.4.2 · STRONG
PAIRED→
SPIKE Prime · #BB02
FW 1.4.0 · WEAK
IDLE→
Daniil's phone
FW — · PHONE
IDLE·
Hub paired. 6 ports detected. Battery 87%.
STEP 02

Save the build

Capture a photo of the actual robot, sketch a port map, and write a short note about what it is supposed to do. This is the context the assistant works from.

  • Photos: top, side, arm detail
  • Port map: roles per port, not just types
  • Intent notes: "follows tape and stops at red"
  • Build copies stay on the iPhone; AI requests share context, and optional feedback can include photos
[ TOP VIEW ]
[ ARM DETAIL ]

Build notes

DRAFT
drivefollowedge-detectarmbeginner-build
STEP 03

Create skills

Ask in plain English. The assistant drafts MicroPython, you read and tweak it, then save the working version into a library you can re-use across builds.

  • Plain-English prompt to MicroPython
  • Skills are diff-friendly text files, not blocks
  • Versions auto-saved on each successful run
  • Library is per-build, sharable as a folder
skills/avoid_obstacle_v2.py● GENERATED
1# Asked: "stop before bumping into anything"
2from spike import PrimeHub, DistanceSensor, Motor
3
4hub = PrimeHub()
5left = Motor("C")
6right = Motor("D")
7eye = DistanceSensor("B")
8
9SAFE_CM = 12
10
11def cruise(speed=40):
12 left.start(speed)
13 right.start(speed)
14 while True:
15 distance = eye.get_distance_cm()
16 if distance is not None and distance < SAFE_CM:
17 left.stop()
18 right.stop()
19 hub.speaker.beep(70, 0.2)
20 return
21
22cruise()
STEP 04

Talk to the robot

Run a skill by voice — "follow the line" — and live sensor state shapes the next bounded action. Motor timing stays deterministic on the hub.

  • Voice intent → skill selection
  • Live telemetry visible during the run
  • Bounded actions: clear stop conditions
  • Failures are explained in plain language
● RUNNING · line_follow_v3t = 14.2s
Left motor
32%
Right motor
28%
Eye L (refl)
68
Eye R (refl)
71
Distance
44 cm
Battery
83%