TraceSpeak Core · The platform Available for pilots
Every signal, understood.
TraceSpeak Core decodes every bus recording, keeps it in a local data lake, finds the anomalies and vehicle states in it, and lets your engineers investigate every signal in the browser or with AI agents. Use it on its own, or as the base of TraceSpeak CT, Fleet and TestDrive.
What it is
Your vehicle data, decoded and searchable.
Point TraceSpeak Core at your bus recordings and network descriptions. It works out what your vehicle network is supposed to do and answers three questions about every recording:
- What happened on the network? Every message and signal, decoded and time-aligned across buses.
- Was it normal? Anomalies found against your network description, without writing a rule.
- What state was the vehicle in? Sleep, wake-up, cranking or partial networking, at every moment.
Everything runs inside your network, on your own hardware.
Five stages
From raw bus traffic to answers.

High-speed bus processing
Skills from writing high-performance network ECU code, applied to CAN, LIN and Ethernet data.
Distributed local data lake
Built for nanosecond timestamps and unlimited scale.
Anomaly detection
Automotive know-how, turned into detection algorithms.
Vehicle state detection
Single events linked into the states every vehicle goes through.
Web and AI investigation
Explore every signal in your browser, or ask the local agents that carry our investigation experience.
Testradamus in Core
Agents that know your vehicle network.
Testradamus is the TraceSpeak family of AI agents. The Core agents work on the vehicle network itself: its description, its anomalies and its signals. Every product adds its own agents on top.
Network description
The Network System Architect reads your DBC, ARXML, LDF and PDX files and works out how ECUs, signals and routes depend on each other.
Anomalies & vehicle states
Every anomaly Core detects is explained: what happened, on which bus and ECU, in which vehicle state, and what it affects.
Investigation
Ask about any signal, message, ECU or bus in plain language: “Why did the gateway stop sending after wake-up?”
How every Testradamus agent works
- Measures with tools, reasons with AI: agents call precise tools for every measurement. The AI reasons over the results; it does not guess numbers.
- Shows its evidence: every answer cites the checks behind it and carries a confidence score.
- Runs locally: an open-weight model on your own GPU, tuned with professional automotive embedded expertise.
Ask Testradamus · illustrative
One signal, from value to impact.
An engineer asks about a single signal. The Core agents answer from the recording and the network description: the value, who sends it, who receives it, and what happens if it fails.
BrakePedal_Pressed was 1 (pressed) at 42.3 s. It travels in PDU_BrakeStatus_01 on Chassis CAN, every 10 ms.
Sender: BrakeCtrlUnit. Receivers: 8 ECUs on 3 buses. CentralGateway routes it to Powertrain CAN and ADAS CAN within 5 ms, where HybridCtrlUnit, InstrumentCluster and the ADAS controllers read it.
If it fails: after 3 missed cycles each receiver switches to its substitute value. Regenerative braking is limited, adaptive cruise control disengages and the instrument cluster shows a brake warning.
Underneath · what the agents checked
- Recording: decoded Chassis CAN at 42.3 s and read the signal value
- Network description: the Network System Architect looked up the PDU, its sender and its cycle time
- Routing: followed the gateway routes to every receiving ECU and bus
- Failure behaviour: read the timeout and substitute value each receiver defines, and the functions that depend on them
Features
Everything between a raw recording and an answer.
Decode & store
Every recording, read and kept in one place.
Every recording, decoded automatically
CAN, CAN FD, LIN, FlexRay, Automotive Ethernet and SysVars recordings (BLF, ASC, MF4) are decoded as soon as they land, together with UDS diagnostics. Nothing to convert, nothing to upload.
One data lake for all of it
A distributed data lake on your own network, with nanosecond timestamps and unlimited scale. Every bus, every recording, aligned on one timeline.
Understand the vehicle
Built from your own network descriptions.
It knows how your vehicle should behave
From your network descriptions (DBC, ARXML, LDF, PDX), TraceSpeak Core builds a model of your vehicle: ECUs, signals, routing, timing, safety relevance and the user functions they support.
Anomalies found without writing a rule
Error frames, bus-off, Ethernet link down, frame drops, silent channels, late or missing messages. Every recording is checked against what the network should do, including failures nobody wrote a test for.
Vehicle states
Events in the context they happened in.
Every state recognised
Sleep, wake-up, cranking and partial networking, detected in every recording, so you know what the vehicle was doing when something went wrong.
Single events, linked
Individual symptoms are tied together into patterns and events, from one signal up to the whole network.
Investigate
Browse it yourself, or ask the agents above.
Every signal in your browser
Explore raw and decoded data from the office or from home, with no tool licences and no files to copy.
Who it’s for
One source of truth for vehicle data.
Network & system engineers
Every bus and every ECU on one timeline, checked against the network description.
Software developers
See how your ECU really behaved on the bus, and ask why, without waiting for someone to send you a log.
Validation engineers
Anomalies found in every recording, not only in the ones someone opened.
Team leads & management
The same signals, anomalies and answers your validation engineers and developers see, so decisions rest on the same data.
Supported inputs
What TraceSpeak Core reads.
| Source | Details | Status |
|---|---|---|
| Bus recordings | CAN, CAN FD, LIN, FlexRay, Automotive Ethernet, SysVars. File formats: BLF, ASC, MF4 | Supported |
| Diagnostics | UDS requests, responses and DTCs | Supported |
| Network descriptions | DBC, ARXML, LDF, PDX | Supported |
See your own data in Core.
We will show TraceSpeak Core on a simulated vehicle, then on your recordings.