Which motor actually wins?
Give two motor pages the same wheel, chassis and pack voltage. Race both simulations over identical ground.
- inputs
- 3 product pages
- output
- evidence-backed winner
S2
Scrape2SimWEBGL · REAL COMPONENTS
Open the labloading real component scene
s2s-real-components-lab.glbwebgl · pbr + image-based lightingaz -34° dist 12.0 rotating
WEB → HARDWARE
Prove the build
before you power it.
Paste real product URLs. Scrape2Sim pulls the BOM with its evidence attached, gates it on electrical limits, simulates the whole machine in MuJoCo — then shows you where the simulation and your ESP32 disagree.
CH 02 — THREE DECISIONS
Start with a real decision.
Each lab answers one question you would otherwise answer by buying the parts and finding out.
Will the electronics survive?
Pack voltage, driver current, channel count and logic level are checked against each other before anything is wired.
- profile
- ESP32 + TB6612
- output
- pass, warn or block
Does reality match the physics?
Run a bounded bench test over USB and put encoder distance beside the simulated distance. Every motor command is confirmed by a human.
- link
- ESP32 over USB
- safety
- confirmed bounded test
CH 03 — TYPED BOUNDARIES
Nothing reaches physics untyped.
A scraped page never jumps straight into a solver. Five stations, five contracts — each one inspectable, each one able to stop the build.
- 01Public pagesthe open web
Public HTTPS product URLs. Nothing classified, nothing trusted yet.
- 02Scrapebright data
Every spec value carries the snippet it came from and how it was obtained.
- 03Validatetrust boundary
Voltage, current, channels and logic level checked against each other.
- 04Simulatemujoco
The validated spec picks an adapter. You get a scrubbable trajectory.
- 05Measureesp32 hardware
Encoder distance lands beside simulated distance. You confirm every command.
POST /v1/designs/{id}/simulateresults stay evidence-backeda human confirms every motor testDownload the workflow source
READY FOR THE WORKBENCH