UCAN Studio

AI

From AI-assisted scripting to autonomous hardware test farms — UCAN Studio unlocks every level.

The uCAN Engineer Agent

In UCAN Studio every operation is also a programmatic tool, and the GUI, the Python SDK/CLI, and AI agents drive the same backend — full operational parity. That lets the uCAN Engineer agent run the entire loop with you: create topologies, author behaviors and tests, run simulations, inspect traces and logs, and propose fixes.

It works in a plan-and-execute cycle with diff preview, so you stay in control while the agent does the heavy lifting — in the GUI today, and headless in CI and on hardware test farms.

Four Levels of AI in Software Engineering

From a programmatic platform with no AI, to assisted scripting, to agents that run the whole loop — even headless on hardware test farms. UCAN Studio is built for every level.

L1

No AI — the platform moat

The unified, programmatic engineering platform — the foundation everything else stands on.

At this level

  • Simulation, test, trace, dev
  • One workbench, all protocols
  • Engineer-driven, manual

The automotive barrier

  • Fragmented toolchain — many formats & languages
  • Proprietary vendor scripting languages
  • Far less model training data than Python

UCAN is the moat

  • CAN / Ethernet / V2G · sim · test · trace
  • Every operation is programmatic
  • The foundation AI builds on
L2

AI-assisted scripting

AI helps you write behaviors, tests, and scripts — but you still drive the workflow.

At this level

  • Copilot writes code in-tool
  • Behaviors, tests, snippets
  • Human runs everything else

The automotive barrier

  • Vendor scripting languages
  • Models barely trained on them
  • Code is not an end-to-end workflow

UCAN goes further

  • Python-native everything
  • Where models are strongest
  • Typed DBC / ODX / FMU accessors for Python
L3

Agentic AI — human in the loop

An agent runs the full develop → simulate → test → analyze → fix loop while you supervise in the GUI.

At this level

  • Agent builds topologies & sims
  • Runs tests, reads traces, fixes
  • Closed loop — you approve

The automotive barrier

  • GUIs are far harder to drive than programmatic operations
  • Sims & topologies are not programmatic
  • Trace analysis is not exposed as tools

UCAN delivers it today

  • 100+ programmatic domain tools
  • Plan / execute with diff preview
  • uCAN Engineer, in the GUI
L4

Headless agentic AI — autonomous test farms

Coming soon

Agents run headless on cloud and on-prem HIL farms — taking tasks, acting, and reporting back.

At this level

  • Tasks via ALM / messaging
  • Dev, test & analysis on HW farms
  • Scheduled & on-demand campaigns
  • Closed loop — notifies results

The automotive barrier

  • Needs real hardware, headless
  • Orchestration + autonomy
  • Results must flow back to ALM / CI
  • No human at a screen

UCAN is built for it

  • Headless runner · SDK · CI-native
  • Hardware-bound buses
  • Headless on Windows & Linux
  • The frontier we are built for

Put the Agent to Work

See the uCAN Engineer build, run, and analyze on your own networks — in the GUI, in CI, and on hardware.