StarGateVR

Cross-match and enrich your datasets with archives spanning the electromagnetic spectrum:

• Gaia DR3 — Positions, photometry, kinematics (including live ADQL queries)
• eROSITA — X-ray observations
• GALEX — Ultraviolet photometry
• 2MASS — Near-infrared (J, H, K)
• WISE/AllWISE — Mid-infrared photometry
• SIMBAD — Spectral types, classifications, named-object resolution
• MOCA — Stellar association and cluster membership

Imported datasets can be cross-matched and enriched against any of these archives. Results are merged by source ID or positional cross-match (typically 3 arcsec) and loaded back into the 3D environment.

Map any combination of parameters to X, Y, and Z axes:

• Physical positions (RA, Dec, distance)
• HR diagram (temperature vs. luminosity)
• Custom spaces — metallicity vs. radial velocity vs. galactic latitude

Visualize up to five dimensions simultaneously:

• X, Y, Z → spatial axes
• Color and brightness → two additional parameters
• U, V, W → Motion animation as a sixth dimension

AI-assisted analysis: 

• Ask the assistant to generate plots — displayed in VR and saved as PDF
• Produce computed columns and HR diagrams with log-scaled axes
• Real-time progress updates for long-running operations

Auto-computed quantities: If your CSV includes RA, Dec, and parallax, StarGateVR automatically derives XYZ coordinates, UVW velocities (with proper motion + radial velocity), absolute magnitudes, and color indices. See the CSV Input Data Format Reference for details.

Select and filter star subsets across multiple parameter spaces:

1. Identify an over-density in XYZ position — possible star cluster
2. Gate those stars with a distinct color
3. Switch to UVW velocity — check for common motion
4. Add an HR diagram with metallicity — refine further

Key capabilities: 

• Multiple simultaneous gates, each with its own color
• Gates preserved across space changes
• AND operations to intersect multiple selections
• Export gated subsets as CSV for follow-up analysis

This is the intended workflow: explore in immersive 3D → identify candidates through iterative gating → export to TOPCAT, astropy, or your own pipelines.

Multiple researchers share a VR session with voice communication and synchronized data.

• Co-navigate the same 3D environment in real time
• StarGateVR taps the voice chat audio stream, so you switch between talking to colleagues and querying the AI without interruption
• When one researcher enriches a dataset, results can be shared with all participants

Supports group analysis, advisor-student walkthroughs, and distributed research collaboration.

Import your data: 

• CSV, VOTable, or FITS-derived table formats
• Load via Meta Link (drag and drop), headset browser download, or SideQuest
• Cross-match and enrich imported data with all integrated archives

Where StarGateVR fits:

Use StarGateVR to explore, identify candidates, and enrich datasets — then export for deeper analysis in TOPCAT, astropy, or your own pipelines.

StarGateVR includes a voice-driven AI Assistant, currently available as an opt-in Early Adopter feature. In the app, tap “AI Assistant — Early Adopter” to reveal the interface.

It works, and we’re actively developing it based on real researcher usage. You may encounter rough edges — that’s expected and intentional. Researchers who use it now are helping shape what it becomes.

The assistant handles three types of requests:

Ask about your data. Query statistics, distributions, or outliers in your currently loaded dataset. Select a star and ask “Tell me about this star” to retrieve catalog information from SIMBAD and other sources.

Enrich with multi-wavelength data. Say “Add X-ray and infrared data for these stars” and the system:

• Queries the relevant archives
• Cross-matches by position or source ID
• Merges the results into your dataset

This is the voice-driven way to do the multi-wavelength cross-matching described above, one spoken request can pull from several archives at once.

Retrieve new datasets by description. Say “Get me the stars in the Hyades cluster” and the system:

• Recognizes this as a named stellar association/cluster membership request
• Looks up the Hyades membership list through MocaDB
• Retrieves the associated Gaia DR3 star data linked to those members
• Loads the resulting star set into VR

You describe what you want in plain language — you never see the underlying queries. The AI classifies your intent automatically. When a request is ambiguous, it asks for clarification.

The AI Assistant never generates scientific data. Every measurement — photometry, positions, spectral types, cross-match results — comes directly from established astronomical archives through deterministic queries.

The LLM’s role:

• Understand what you asked
• Route it to the correct database
• Return archive results — or an explicit null if there’s no match

Where the assistant generates text (summarizing results, suggesting interpretations), the underlying data is always visible in the 3D environment for direct inspection. The system is designed around one distinction: “looked it up” vs. “made it up.”