Agent-native measurement lab

The vehicle data lab your AI agent can run

MDX isn't just another oscilloscope. It exposes your MDF4 and vehicle/ECU data through a built-in MCP server — so Claude Code, Codex, and Gemini load logs, query thousands of channels, plot signals, and find events by chat. And when you want to drive it yourself, it's a full oscilloscope-grade lab. Offline-first.

Speaks MCP — works with
Claude Code Codex Gemini Claude Desktop VS Code
hub AI agent ↔ mdx · mcp localhost:8000/sse
Load run_42.mf4 and chart coolant temp vs. load
boltupload_log → list_signals → plot_signals
coolant_T vs load2,531 ch
Charted over 1,204 s. Coolant peaks at 96°C under load — no overshoot. Run a misfire scan next?
format_list_bulleted

Universal Formats

Native support for MDF, MF4, CSV, and Excel. The only workspace you need for measure data analysis.

bolt

Insane Speed

High-performance LTTB downsampling and sub-second channel rendering for GB-sized files.

cloud_off

Offline-First

Total privacy. Run backend and frontend locally with zero data ever leaving your machine.

devices

Any Deployment

Run the Next.js web app against your local API, or install the native Vehicle Lab desktop suite — a single self-contained build with the analysis backend bundled in.

Professional Lab Toolkit

Engineered for calibration engineers and data scientists who require millisecond precision and deep signal insights.

data_object

Data & File Handling

Native support for ASAM MDF v3 and v4, plus CSV and Excel. Large, multi-gigabyte files load through memory-mapped reading to keep RAM low, and every signal is automatically extracted and indexed on import — ready to query instantly.

Variable Explorer
travel_explore

Variable Explorer

Instant global search and filtering across every signal in your file, with intelligent matching that resolves naming differences across DBC databases and OEMs. Drag and drop straight to instruments — no configuration overhead.

Oscilloscope
show_chart

Professional Instruments

Interactive, high-resolution oscilloscope with dual A/B cursors, resizable panels, and time-aligned multi-signal tables. LTTB downsampling keeps even dense traces smooth for precise event correlation and value tracking.

psychology

Advanced Analytics

A full diagnostics suite — engine misfire detection, gear-hunting analysis, fuel and BSFC metrics, and WLTP cycle analysis. Generate empirical maps with Kriging, RBF, or spline interpolation, rendered as 2D heatmaps and 3D surfaces.

compare_arrows Multi-file analysis

Two files? Two thousand? Compare them properly.

A bench oscilloscope shows one capture at a time. MDX lines up entire test campaigns — end-to-end or side-by-side — and aligns them for you.

serial · one timeline
log_01
log_02
log_03
0 st →
merge

Serial

Concatenate time. Stitch logs from a long trial — or a fleet logged back-to-back — into one continuous timeline. Scrub, cursor, and measure straight across file boundaries as if it were a single recording.

Long trials · sequential fleet runs
parallel · overlaid
sync
−20°C 25°C 40°C
layers

Parallel

Overlay & auto-align. Put the same signal from many files on one axis — different temperatures, vehicles, or idle points — then pick a sync signal and MDX time-aligns every run automatically. The comparison a bench oscilloscope simply can't do.

A/B/N comparison · sync-signal auto-align
smart_toy Built-in AI

Ask your data anything with DiagAI

DiagAI is the analysis assistant built into MDX. Ask questions in plain English and get engineering-grade answers — signal detection, statistics, correlation, and anomaly checks — without manually wiring up a single plot.

  • forum Natural-language queries — "average RPM during acceleration", conditional filters, and multi-signal comparisons, answered instantly.
  • groups Multi-agent diagnostics — specialized agents for misfire, gear hunting, fuel, drivability, and WLTP coordinate on a single request.
  • shield_lock Private by design — run against your own local model so data never leaves your infrastructure, with isolated, per-OEM knowledge bases.
  • alt_route Smart model routing — routes each query to the right model for the best balance of speed and depth, with automatic failover.
smart_toy DiagAI local model
Show me data where rpm < 2000 & torque > 150

Found 4,812 matching samples across 2 signals. Mean torque 168 Nm, peak at 1,840 rpm.

Detected one anomaly window near t=312s — open misfire analysis?

Ask about your file…
send

Built to Fit Your Pipeline

MDX ships as a native desktop suite and exposes everything it does to your own scripts and tools.

desktop_windows

Native Desktop Suite

Installs as a self-contained Vehicle Lab desktop app — a single native build (Tauri/Rust) with the analysis backend bundled in. No server to host, no runtime to manage, fully offline.

map

GPS & Route Mapping

Plot drive routes on offline map tiles and correlate any signal against location — no internet connection required, so field and test data never leaves the lab.

terminal

Python SDK & REST API

Automate everything from your own scripts. A documented REST API and Python SDK expose files, channel metadata, signal data, batch analysis, and anomaly detection.

hub Model Context Protocol

Drive MDX from Claude Code, Codex or Gemini

MDX ships a built-in MCP server, so your AI coding agent can operate the lab directly. Connect once to a local endpoint and ask in plain language — the agent uploads logs, queries thousands of channels, plots signals, and detects events, all against data that never leaves your machine.

  • build_circle 12 native tools — health checks, log management, signal stats, plotting, event & similar-event detection, and legacy config import.
  • cable One-line connection — point any MCP client at the local SSE endpoint. No API keys, no cloud round-trip.
  • diversity_2 Works with your agent — Claude Desktop, Claude Code, VS Code, Codex, and Gemini all speak the same protocol.
  • lock Offline & private — the server wraps your local analysis backend; measurement data stays on-prem.
claude_desktop_config.json
// connect any MCP client — one block
{
  "mcpServers": {
    "mdx": { "url": "http://localhost:8000/sse" }
  }
}
// VS Code / Codex: { "servers": { "mdx":
//   { "type": "sse", "url": ".../sse" } } }
Plot vehicle speed and find hard-braking events in log 3

boltmdx · plot_signals, find_events

Plotted Veh_v over 1,204 s. Found 7 hard-braking events (decel > 4 m/s²) — largest at t=512 s. Chart + event table attached.

MDX vs Legacy Desktop Tools

The familiar workflow of a legacy desktop analyzer — plus the comparison, mapping, and AI control a bench oscilloscope can't touch.

Feature Legacy Tools (ETAS MDA) MDX Lab
Large File Performance Slow loading / High RAM Sub-second loading / Memory Mapped
UI Modernity Legacy Desktop Widgets Modern Web-first Responsive
Deployment Flexibility Windows Desktop Only Web, Desktop, or Portable EXE
Automation / Scripting Proprietary API Python / JS / REST Native
AI Diagnostics Manual analysis only Built-in DiagAI assistant, private-deployable
Multi-file Comparison Manual overlay, no alignment Serial concat + parallel overlay with sync-signal auto-align
Calculated Signals Basic formula editor Server-side engine — accumulators, windowed stats, filters, latches
GPS & Location Not built-in Offline route mapping on the timeline
AI-Agent Control None MCP — drive from Claude Code, Codex, Gemini
Config Compatibility Native configs only Imports your existing .xdx/.xdt configs

Ready to accelerate your vehicle data analysis?

Join hundreds of calibration engineers switching to MDX for a faster, more flexible, and modern data laboratory environment.

FAQ

Frequently asked questions

Can MDX be operated by an AI agent or LLM like Claude Code, Codex or Gemini?expand_more

Yes. MDX ships a built-in MCP (Model Context Protocol) server exposing 12 tools — health checks, log management, signal statistics, plotting, event and similar-event detection, and legacy config import. Any MCP client (Claude Code, Claude Desktop, Codex, VS Code, Gemini) connects to the local SSE endpoint with a single config block, then loads logs, queries channels, plots signals and finds events from a chat prompt. No API keys and no cloud round-trip; measurement data stays on the local machine.

What measurement file formats does MDX support?expand_more

MDX natively supports ASAM MDF v3 and v4 (.mdf, .mf4), CSV, and Excel. MDF files are read via memory-mapped I/O so multi-gigabyte files open without loading the full dataset into memory. Signal extraction and indexing happen automatically on import.

Can MDX work fully offline without an internet connection?expand_more

Yes. MDX is offline-first. The desktop application (packaged via Tauri) bundles its own backend and ships with offline map tiles for GPS route visualization. No data leaves the local machine unless you explicitly configure an external AI endpoint.

Does MDX provide an API for automation and batch processing?expand_more

Yes. MDX exposes a documented REST API (running locally at 127.0.0.1:5001) and ships a Python SDK (mdx_client) for scripting signal queries, batch analysis jobs, anomaly detection, and map tile operations. The SDK mirrors the REST API one-to-one.

What analytics does MDX provide beyond basic signal viewing?expand_more

MDX includes engine misfire detection, gear-hunting analysis, fuel and BSFC calculations, WLTP cycle analysis, and empirical calibration map generation using Kriging, RBF, and spline interpolation. Maps render as interactive 2D heatmaps or 3D surfaces. LTTB downsampling keeps rendering fast even for high-frequency signals.