Guided demo

See Maestro in four steps

We'll follow one real test — the Artifact Visualization Demo, a frequency-response sweep — from how it's built, to running it, to reading the report, to asking questions of the data. About two minutes.

Tests are authored by AI

Walkthrough You describe intent; your AI agent writes the test.

You don't open a sequence editor or hand-write code. You point your own AI agent — Claude, or any MCP-capable assistant — at Maestro's MCP server, then say what to measure and the limits in plain language. It generates the test as readable, version-controlled YAML plus any Python and/or C# methods it needs. You review the intent, not the source.

To be clear about which AI is which: this one is your agent, on your machine, driving Maestro over MCP — that's why this step is a walkthrough rather than something you can click here. The chat assistant in steps 2–4 is the same idea, hosted in this demo (bottom-right corner) so you can try it without installing anything.

You say

“Sweep the Bode response of the amplifier. Check the gain at 1 kHz is between −10 and −4 dB and the phase is −100° to −75°. Capture an interactive Bode chart and export the sweep as data.”

Maestro generates ↓

It writes — you typically never touch this

test.yaml

steps:
  - name: "Bode Sweep — Gain at 1 kHz"
    type: mock
    measurement:
      name: "GAIN_1KHZ"
      value: "{{test.gain_1k}}"
      low_limit: -10.0
      high_limit: -4.0
      unit: "dB"

  - name: "Generate Interactive Bode Chart"
    runner: python
    parameters:
      module: "bode_artifacts"
      function: "generate_bode_plotly"

bode_artifacts.py

def generate_bode_plotly(frequencies, gains_db, phases_deg):
    """Emit an interactive Plotly chart artifact."""
    fig = make_bode_figure(frequencies, gains_db, phases_deg)
    with open(_artifact_path("bode.plotly.json"), "w") as f:
        f.write(fig.to_json())     # Maestro renders it inline

Now run it →

Run it and watch live

Live Two ways — do it yourself, or ask the chat assistant.

The Test Runner streams progress, variables and each step's verdict in real time. We've pre-selected the Artifact Visualization Demo for you — just press Start Test.

Open it in the Test Runner or

Ask the chat assistant to run it

Run the Artifact Visualization Demo test on serial DEMO-001 and tell me the verdict.

Paste it into the chat drawer (bottom-right) and send.

See the report →

Read the full report

Live Every measurement is queryable data — not a log file.

Each run produces a report with the pass/fail verdict, a measurement table against its limits, and any artifacts rendered inline — including the interactive Plotly Bode chart (zoom, hover, export) and the sweep data as a table. Open Test Results, search Artifact, and open the most recent run.

Browse Test Results

Ask questions of the data →

Ask questions of your results

Live Plain-language questions over every measurement.

Now zoom out — the real power is querying across every run on the floor, not just the one you launched. Because results are structured data, the chat assistant reasons over yields, process capability (Cpk), marginal passes and failure patterns. Try one in the chat drawer (bottom-right):

Which product or test has the most failures, and what's driving them?

Find a measurement with real spread and show me its distribution across all runs.

What's the overall pass/fail yield across all test results?

Or explore Process Capability Or filter results by hand

That's the loop: author with AI → run → report → query. Everything you saw is one connected system.