Drag to explore ↔RGBDEPTHPIXELS → DISTANCE
Look through the image.
Same camera. Same tick. Appearance and geometry, aligned.
01 / EXPERIENCE THE DATAOne moment.
One moment.
Every dimension.
This is real data, not a product mockup. Switch layers, reveal the signals, and follow the decisions behind a ten-second encounter.
MESOCORT / MULTIMODAL EXPLORERDUST II · ROUND 01 · PLAYER 06
ONE PLAYER. ALIGNED PERCEPTION AND ACTION.SYNCHRONIZED SAMPLE
01 / RECORDED INPUTS Live actions
W
ASD
SHIFTCTRLSPACE
◉ PRIMARY FIRE◉ SECONDARY
R RELOADE USEZOOMINSPECT
Waiting for input data…
02 / PLAYER STATE TICK 5380
03 / INPUT SEQUENCE & MOVEMENTEVERY ACTION, IN SEQUENCE
Drag to orbit · scroll to zoom · arrow keys supported
CT T Navigation network
PREVIEW WITH HUD1080p / 64 fps
OBSERVER / PLAYER 06
TARGET EXPOSURE ENGINE OUTPUT
—measured visibility
Body regions shown as an anatomical reference. The real percentage is target-level; this export does not contain separate region masks.
FRONT ↑ Player-relative sound field
Drag to reveal depth ↔RGBGEOMETRY DEPTH
NearFarMetric source · static map geometry
ONE TIMELINE → THREE CONNECTED VIEWSPLAYER 06 / FIRST PERSON
PLAYER VIEW RGB
GEOMETRY DEPTH
WORLD CONTEXT 2D
TICK 5380
Observation RGB + depth + available cues→Human action Keys + aim + fire→Next state Grounded in the next tick
02.56 / 10.00 sTICK 5380
0 s2 s4 s6 s8 s10 s
One real encounter. Ten players. Every layer connected by the same tick. Player labels are neutral; original video retains the in-game HUD.Go deeper in the full explorer ↗
02 / LOOK CLOSERThere is more here
There is more here
than meets the eye.
Lift the image into 3D. Reveal what a player could hear. Turn a path through the map into a sequence your model can learn from.
Drag to explore ↔RGBDEPTHSPACE + MOVEMENT
POSITIONS → CONTEXT
Every move has a place.
Navigation, named areas and the paths people actually take.
OUR 3D EXPLORERA RECORDING → A WORLD
Step outside the camera.
Explore the scene in 3D, connect movement to geometry, and inspect an encounter from another angle.
A PLAYER’S SOUND FIELD
SOUND → AVAILABLE INFORMATION
Hear beyond the frame.
Who made the sound, where it came from, and who could hear it.
RGB / depth, map and sound previews share tick 5380. The 3D image shows a separate view in our full explorer. Depth represents the static environment.
03 / THE DIFFERENCE IS THE CONTEXTA video shows the moment.
A video shows the moment.
Connected data opens it up.
Start with the human experience. Add the structure, perspectives and outputs your research needs.
| What you can work with | A standalone videoThe visible experience | Mesocort connected dataThe experience, decoded | Your tailored datasetShaped around your question |
|---|---|---|---|
| Observation | Recorded pixels | RGB + camera + aligned depth | Your views, resolution & HUD choice |
| Human action | Inferred from the image | Recorded inputs, aim & player state | Selected behaviors & time windows |
| World context | What is in the frame | 3D geometry, navigation & trajectories | Spatial features for your task |
| Perception | What is visible or audible in the recording | Per-target visibility & listener-level sound | Observer-specific cues & enrichments |
| Ready to use | A media file | Layers linked by player and tick | Selected, aligned & validated exports |
64 HzOne observation every 15.625 msPerception → action → next state
Explore all seven data layersInputs · space · video · visibility · sound · depth · egocentric +
Comparison with a standalone video file, not a claim about another provider. Corpus scale and generation capabilities supplied by Rankacy; layers are generated on demand, subject to source coverage.
04 / THE DATA. THE TOOLS. THE EXPERTISE.Your question.
Your question.
Your version of the data.
We build the infrastructure around the dataset, too. We can select, transform and enrich the data, then deliver the views and features your team needs.
Select
Choose the situations, players, perspectives and time windows that matter.
A focused slice of human experienceEnrich
Add visibility, spatial sound, geometry and features tailored to your question.
More context around every actionRender
Generate aligned RGB and depth. Choose the camera, resolution and HUD.
The observation your model needsValidate & deliver
Inspect alignment, compare outputs and assemble an export for your workflow.
Data you can inspect and build onBUILT AROUND YOUR RESEARCHBring a task.
Bring a task.
We can build the data around it.
From a focused experiment to a custom training corpus, we work with you on the right selection, enrichments and delivery.
Build with Mesocort ↗01 / WORLD MODELSRGB + depth + action → next state
Aligned sequences, camera records and human inputs.
02 / PERCEPTION RESEARCHWhat this player could see and hear
Observer-specific visibility and sound context.
03 / BEHAVIOR MODELINGSelected encounters + enriched telemetry
Movement, aim, timing and the context around a decision.
05 / BUILT ON THIS FOUNDATIONThese data already
These data already
power intelligence.
From a rich record of human play to models that explain what changes a game. This is already a working product.
LIVE AT RANKACY LABS
GPRT Pro
From what happened
to what made the difference.
Our model turns match state into live win probabilities, identifies decisive moments, and explains the events behind a round’s outcome.
01 Live win probabilities02 Impact moments & player analysis03 Outcome explanations & summaries04 Conversational match exploration
Experience GPRT Pro ↗MATCH INTELLIGENCE / PRODUCT WORKFLOW
GPRT ProCONTEXT → PREDICTION → EXPLANATIONWorkflow illustration. Open the live demo for real model outputs.
GAME STATEPLAYER ACTIONSROUND CONTEXT
GPRT ProCONTEXT → PREDICTION → EXPLANATION01Predict
How does the chance of winning evolve?
02Explain
Which moments changed the outcome?
03Explore
What did each player contribute?
A model family, built for different questions. GPRT Flash supports fast inference. GPRT Pro goes deeper into match analysis.
Explore Rankacy Labs ↗06 / WHAT COMES NEXTA playground for humans.
A playground for humans.
A foundation for AI.
A foundation for models that understand worlds, anticipate actions and learn from human experience. Explore the larger possibilities.
Research horizons, alongside the products already built above. Applying models beyond games remains a separate research and validation challenge.
07 / OUR OWN RESEARCH AT MESOCORTWe build with this data.
We build with this data.
And ask what comes next.
We are Mesocort. Alongside the dataset and tools, we conduct our own research into human behavior and the intelligence behind a decision.
HUMAN BRAIN / ANATOMICAL ATLAS
Loading 3D brain…
01 OrientSuperior colliculus→02 RelayPulvinar · thalamus→03 RespondAmygdala
A proposed visual route involved in rapid responses to threat. The dimmed thalamus and brainstem provide anatomical context.
Select a region to highlight it.Drag to rotate · pinch to zoom
MRI-derived cortical anatomy. Region colors identify anatomy, not measured brain activity.
MESOCORT RESEARCHFrom human play
Explore our research ↗
From human play
to human behavior.
We use this rich record of perception and action to study how people respond under pressure. Our models Ember and Orbit investigate fight-or-flight outcomes from behavioral telemetry.
THE QUESTION WE ARE EXPLORING
Can the moments before a decision help us predict how a person will respond?
EMBER
Behavioral prediction
ORBIT
Behavioral prediction
Ongoing research. Our paper is under review; the methodology is available under NDA through Mesocort.
REAL HUMAN EXPERIENCE→OUR ENRICHMENT TOOLS→OUR MODELS & RESEARCH
Anatomy: FreeSurfer fsaverage5 + Desikan–Killiany regions, via PennLINC · Source & credits. Visual pathway research: Kragel et al., 2021 · McFadyen et al., 2019.
7 MILLION GAMES. YOUR NEXT BREAKTHROUGH.
Bring the question.
We bring the experience.
A dataset. A toolchain. A research partner.
Let’s shape the data around what you want to discover.