Loading...
Server Online

Multi-Modal Cognitive Intake

Search HSK materials and concepts using typed text, voice microphone input, or image file uploads.

Response Time: --

Search Query Results

Enter a query, click the mic, or upload an image to view HSK results.

Live Feed & Companion Scanner

Stream live webcam capture or share a screen window to run real-time Edge AI text/object translation.

Offline
0.00 % Spatial Var
0 ms Dwell Lock
Edge Tier
Frame Capture-- ms
Dwell Lock-- ms
Edge AI inference (OCR/Object)-- ms
TTS Vocoder Synthesis-- ms
Total Perception Latency-- ms

🚀 Cross-Tab Scanner Companion

Drag this button to your browser bookmarks bar. Open any website in another tab, click the bookmark, and highlight text or drag crop boxes to translate in real-time!

MNee Web Scanner
Video Intake & Perception Events
[System] Video pipelines inactive. Enable Video Intake to start.

WebRTC Audio Streaming

Establish a low-latency WebRTC connection to stream microphone input and evaluate real-time hybrid VAD barge-ins.

Disconnected

Barge-in Simulation

Pipeline Telemetry & Logs

-- ms Network RTT
-- ms VAD Loop Time
Live Telemetry Event Logs
[System] Ready to establish WebRTC connection.

Learner Profile Setup

DIHCL ZPD Item Selector

Retrieve items targeted to the learner's Zone of Proximal Development based on current skill masteries.

BKT Skill Masteries

Real-time Bayesian Knowledge Tracing probability of mastery $P(L)$ values.

Please initialize or load a learner profile.

taVNS & CUTE Haptics Mock Console

Visualizes output waveforms and channel coordinates generated in real time to destabilize L1 attractors.

taVNS Active
Frequency: 30 Hz
Amplitude: 0.0 mA
Timing Alignment: --

32-Channel Haptic Activation Matrix

Wristbands (Pitch Height $F_0$)

Ch 0 - 15 (0-7 Left, 8-15 Right)

Anklets (Pitch Speed $|dF_0/dt|$)

Ch 16 - 31 (16-23 Left, 24-31 Right)

Sine-Wave Vocoder (DSP Test)

Strip vocal traits and test target language-specific spectral configurations.

Pre-Phonetic taVNS Sync Log (Target: 50ms)
[System] Awaiting hardware triggers...

Linguistic Knowledge Graph Subgraph

Query topic concepts to retrieve the local semantic subgraphs injected into prompt contexts.

LMM Grounding Text (Formatted Prompt Context)

Linguistic Graph Nodes & Relations

Active Subgraph Nodes

No active nodes queried.

Active Subgraph Edges

No active relationships.

CI/CD Prompt Evaluation Suite Console

Execute automated prompt compliance suites evaluating schema validity, persona guidelines, and semantic drift checks.

Harness Idle
Prompt Evaluator CI/CD Stdout Console
[System] Evaluator harness ready to compile. Click run to generate test reports.

HSK 1-3 Data Gathering Pipeline

Triggers the parsing engine to harvest lexical taxonomies, radicals, mnemonics, and grammar rules from the design doc, generating visual cards and synthesising vocoded audio.

0 Words
0 Phrases
0 Grammar Rules
0 Scenarios
Awaiting trigger
Data Gatherer Output Logs
[System] Ready. Database will be cleared and rebuilt on run.

On-Device Vector Database Search

Performs sub-millisecond local semantic similarity queries. Type concepts or English words to test retrieval.

DB Search: --  |  Total: --

Query Results

Type a search query to view matching HSK materials.

Live Camera Intake & Bounding Focus Box

Ingest uncompressed stream buffers, filter saccades with a dwell-time gatekeeper, and perform on-device WebAssembly AI.

Offline

Edge vs Cloud Multi-Tier Routing Telemetry

0.00 % Spatial Variance
0 ms Dwell Lock
Edge Decided Tier

Selecting a simulated object bypasses real camera AI and injects a template target image to test BKT, KG routing, and audio contour synthesis.

Sub-500ms Eye-to-Ear Latency Budget

Frame Grab & Processing-- ms
Dwell Stable Lock-- ms
Dual Inference (OCR/Object)-- ms
TTS Vocoder Synthesis-- ms
Total Perception Latency-- ms
Video Intake & Perception Events
[System] Video pipelines inactive. Enable Video Intake to start.