Features

One tuned pipeline,
end to end

Every frame moves through four stages — detection finds the object, tracking holds it, color fusion confirms it, and your hardware acts on it. Every stage is yours to tune from a clean local dashboard, in real time.

Interactive Ecosystem Playground

Experience the platform live

Click between real-time vision servoing, dataset labeling with SentinelLabeler, empirical silicon benchmarks, and zero-GC developer APIs.

Live Interactive Studio Sentinel Core Vision v1.7.0
1,000Hz Output Locked
INGEST: 240 FPS (DXGI) INFERENCE: 2.9 ms (FP16) TIMER JITTER: 0.18 ms ACTIVE TARGET: #104 (ADS LOCKED)
Ingest Rate
240 FPS
GPU Latency
2.9 ms
Hardware Bus
1,000 Hz
Timer Slew
0.18 ms
Kinematic Visual Servoing IMM FILTER ON

Time-optimal phase-plane snapping automatically leads moving targets without overshoot or jitter. Click any target on the canvas to switch lock.

Interactive real-time preview • Zero CPU allocations
SAM 2 • 1-CLICK POLYGON SEGMENTATION
Click anywhere on the object to reposition SAM 2 prompt point
Dataset Annotation Engine

From Raw Capture to YOLO in Minutes

Built specifically for capture card video streams. SentinelLabeler eliminates tedious manual frame-by-frame labeling with cutting-edge AI assistance:

⚡

Segment Anything 2 (SAM 2)

Click once on any target to instantly generate tight polygon segmentation masks with high edge fidelity.

⏩

Temporal Frame Propagation

Annotate a single frame and propagate bounding boxes across up to 300 sequential video frames automatically.

📦

Direct YOLOv8 & YOLOv11 Export

Outputs clean normalized YOLO text labels and train/val directories ready for direct PyTorch / ONNX training.

2.9 ms
Mean Neural Inference Latency
Tier 1 · Esports Grade (144Hz–240Hz+ Ready)
Architecture NVIDIA GeForce RTX 5050 Laptop GPU
Execution Provider TensorRT / DirectML (FP16 Engine)
Peak Throughput 345 FPS
Video Capture 240 FPS (DXGI Desktop Duplication)
Timer Jitter 0.32 ms (Multimedia 1000Hz)
Download Free Benchmark Tool → 105MB Portable • 0 Dependencies
// ============================================================================
// Sentinel Core Vision — Zero-GC 1,000Hz Hardware Dispatch
// High-Frequency Visual Servoing & Kinematic State Estimation
// ============================================================================

[StructLayout(LayoutKind.Sequential, Pack = 1)]
public readonly record struct DirectHidPacket(
    byte ReportId,       // 0x01 = Direct Controller Emulation
    short StickLeftX,    // [-32768, 32767] Normalized Deflection
    short StickLeftY,
    short StickRightX,   // Servoing Saccade Step
    short StickRightY,
    uint ButtonsMask,    // Raw Gamepad Triggers & ADS
    uint TimestampMicro  // 1,000Hz Jitter Compensation
);

public sealed class ServoKinematicPipeline : IDisposable
{
    private readonly OutputLoop _outputLoop;
    private readonly ProjectionMagnetCleanAimCore _aimCore;

    [MethodImpl(MethodImplOptions.AggressiveInlining | MethodImplOptions.AggressiveOptimization)]
    public void DispatchServoTick(in ReadOnlySpan targets, in VideoFrameHeader frame)
    {
        // Zero allocation pass: IMM filter estimates target velocity & acceleration
        ref readonly var activeTarget = ref _aimCore.EvaluateBestTarget(targets, frame.Timestamp);
        
        // Time-optimal phase-plane snapping with anti-deadzone compensation
        Vector2 stickDeflection = _aimCore.ComputeKinematicStep(activeTarget);
        
        // Zero-GC atomic transmission to Titan Two hardware via USB DirectHID
        _outputLoop.EnqueueStickDeflection(stickDeflection.X, stickDeflection.Y);
    }
}
Detection

Real-time detection, every frame

A fast ONNX vision model scans every captured frame with GPU acceleration across multiple execution providers — CUDA, TensorRT, DirectML, or OpenVINO — and surfaces the result on a status-first dashboard so you always know the engine is healthy. Because it runs locally, it's designed for low-latency processing, and you read the live system state at a glance instead of guessing.

  • Multi execution provider: CUDA / TensorRT / DirectML / OpenVINO
  • Live system health & output-rate readout
  • Hot-swap models without restarting
localhost:5000 — Dashboard
Sentinel Core Vision dashboard showing live detection status and tracking tuning
Vision

Inspect every decision

The Vision page puts the raw feed and the AI overlay together — detections, tracking, ownership, and color, all drawn live as the engine decides. It turns a black box into a window, so when you change a setting you can watch the effect instead of trusting it blind.

  • Live detection + tracking overlay
  • Searchable class filter & object targeting
  • Feed-first layout — debug layers one tap away
Vision · live truth
Live Vision view with the AI detection and tracking overlay
Color fusion

Lock the right object by color

A live color scout ranks the colors in your frame so the object outline is obvious, then HSV color confirmation runs alongside the model to agree on the same object. It keeps you locked on the object of interest and off any ignored class — and the sticky preview sits right beside the controls so you tune it by eye.

  • One-click "apply object color"
  • Excluded-color safety & consensus gate
  • Export & import your exact setup
Color detection
Color detection page with live preview and color scout
Tuning

Tune from simple to developer

Start with a handful of plain-language sliders for feel, then reveal advanced and full developer surfaces only when you want them. Nothing lives in a config file — every parameter is a control in the browser, so you can go as deep as you like without ever leaving the dashboard.

  • Simple · Advanced · Developer disclosure
  • Every parameter is a labeled control
  • Save & load full profiles in seconds
Developer tuning
Developer tuning surface with the full set of engine parameters
Hardware output

Connect your setup, output cleanly

Setup walks you through your capture source and output device and checks the whole chain is ready before you start. Output goes through a supported output device so it reads like native input — no drivers injected into your application — which keeps your system clean and your input native.

  • Guided capture + output device setup
  • Native input via a supported output device
  • Readiness check before you start
Setup · connect your system
Setup page for connecting capture source and hardware output
Why not just do it by hand?

Calibration, not guesswork

The usual alternative is a hand-written script and a lot of trial and error. Here's the difference.

Manual scripts & trial-and-error

  • ✕ Guess a value, test, guess again
  • ✕ No way back once you overwrite a setting
  • ✕ Recoil patterns measured by eye
  • ✕ A script that only works for one weapon

Sentinel Core Vision

  • ✓ Try a candidate setting, compare, decide
  • ✓ Roll back any candidate in one click
  • ✓ Recoil patterns measured automatically
  • ✓ Per-weapon profiles, saved and reusable
And the rest

More in the box

The details that make it feel finished — and keep everything on your own machine.

Updated September 24, 2026

▣

Four aim engines to choose from

Kinetic, Glide, Magnet, and Obsidian — each with its own color identity in the dashboard and its own tuning layout. Pick the one that matches how you play.

Kinetic & Glide production-ready · Magnet & Obsidian advanced
↺

Calibration you can undo

Try a candidate setting, keep it if it feels better, or roll it back in one click. Export a support bundle if something still looks off.

Candidate → keep or rollback
⚡

Flash Titan Two from the dashboard

Prepare and flash GPC scripts, read slot state, and watch live Titan telemetry — no separate Titan software required.

USB HID, built in
⇄

No Titan Two? No problem

Run Sentinel through a virtual Xbox controller (ViGEm) or directly through an existing Xbox Remote Play session — no extra hardware required.

Two more supported output paths
⌕

Tell it what's wrong, not what to change

Jittery, overshooting, grabbing the wrong target, too sticky, dead zone too tight — pick the symptom and the dashboard points you to the right control, plus what to check next.

Symptom-based tuning picker
⟳

Hot-swap models

Load a different ONNX model on the fly — no restart, no downtime. Compare detection models in seconds.

Drop-in model manager
⛁

Save & load profiles

Capture an entire tuned setup as a profile and switch between them per source or per feel in one click.

Export & share your config
⟿

Multiple swappable tracking engines

Choose how tracking behaves — from a bare baseline to predictive, belief-following controllers — live.

Per-engine attachments & memory
◎

FOV & zone control

Define where the engine is allowed to look and act, with adjustable field-of-view and engagement zones.

Profile-aware FOV rings
⬓

Capture-card native

Works from a real capture feed at the device's native resolution and refresh — no special drivers required.

Auto-selects max native mode
⚿

100% on-device

The engine runs entirely on your PC. The only network call verifies your license — never your activity.

No telemetry on your activity
◈

Native Studio app

A native Windows companion app launches and manages the engine, with native WinRT capture and an output loop tunable up to 1000Hz.

WinRT capture · 1000Hz output

Ready to tune your own?

Activate in one click and dial in your first profile in minutes — every stage, in your browser.