01 / Motion Control
DetectingGimbal Position
Adjust pan and tilt to move the robot smoothly through robot.motion.move_to().
00 / System Overview
One bench to validate the whole robot.
From motion and lighting to camera and full-duplex audio, every test uses public SDK APIs from the current Python Application and runs through its authorized Device channel.
System Idle
00.5 / Device Access
Connect Robot
Enter the six-digit pairing code shown on the device. The local SDK Daemon handles pairing, and the code is never stored in Test Bench logs.
Device—
Firmware—
Capabilities00
Last Sync—
00.8 / Resource Lifecycle
Device Resource Monitor
Continuously monitor internal memory, largest contiguous blocks, and media resources; compare against the startup baseline after RTC stops.
Lifecycle Verdict
Waiting for Device Snapshot
No evt.sdk.resource_snapshot received
- Internal Free
- —
- Largest Block
- —
- DMA Free
- —
- DMA Largest Block
- —
- PSRAM Free
- —
- PSRAM Largest Block
- —
- Boot Minimum Internal RAM
- —
OwnersWaiting for Telemetry
Against BaselineWaiting for Baseline
Release ResultNo RTC Stop Record
Recent Feature Resource Snapshots
Worker stack sizes describe resident allocations, not total memory owned by a feature. Full transition history is printed to the device log.
| Stage | Internal Free | Largest Block | DMA Largest Block | TTS Stack | SFX Stack |
|---|
02 / Lighting Control
DetectingBody Lighting
Set color, brightness, zones, or looping effects through the public robot.lights API.
02.5 / Screen Animation
DetectingConcurrent Animation Validation
The device reports the complete animation catalog available on its SD card. You can switch animations manually or use randomized playback during an audio/video session.
Waiting for the device animation catalog
02.8 / Edge Vision
Face Tracking Test
Run tracking on the robot, with optional matching preview frames. Starting tracking moves the gimbal; stopping holds its current position. Stop live video before starting tracking.
Waiting for device capabilities
Awaiting Test
Checking face preview support
—
02.9 / Model Inference
Gesture and Face Model Test
Test gesture and face detection with optional preview and no gimbal movement. Stop camera features before starting.
Awaiting Test
—
03 / Live Video
DetectingLive Camera Preview
The browser receives Watcher MJPEG video directly over a dedicated LAN channel; full-duplex audio continues over WebRTC, and the Python Application does not relay media frames.
- State
- IDLE
- Display FPS
- 0.0 FPS
- Source / Target / Sent
- —
- Average Frame / Send P95
- —
- Congestion Pressure
- Browser 0 / Animation 0
- Animation Actual / Target
- No active animation detected
- Resolution
- —
- Local Drops
- 0
Waiting for Video
WATCHER-RTC / MJPEG-DATA
04 / RTC Audio
DetectingFull-duplex Audio Call
Stream the computer microphone to the Watcher speaker while playing the Watcher microphone in the browser. The current Python Application handles signaling while audio uses a direct LAN WebRTC connection.
- Session State
- IDLE
- Computer Microphone
- Available
- Computer → Watcher
- 0
- Watcher → Computer
- 0
- Robot Microphone Capture
- 0
- Robot Packets Sent
- 0
- Physical Mic Peak / Browser Playback
- 0 / 0.000
- Device Echo Cancellation
- Waiting for Device Telemetry
- Real-time Audio Stage Latency
- Waiting for Stage Latency Samples
- Validation Action
- Speak toward the robot
05 / Audio Downlink
DetectingSpeaker Stream
Stream the built-in 24 kHz mono PCM sample through the Application channel.
06 / Image Capture
DetectingCamera Capture
Capture one JPEG photo and return it as a managed Application artifact.
No Image
Capture channel ready
07 / Audio Uplink
DetectingMicrophone Recording
Receive Watcher Opus audio, decode it to PCM through the SDK, and save a validated WAV file.
5 s
08 / Capability Contract
Device Capability Matrix
09 / Runtime Trace