TheStarDB is a living stellar encyclopedia assembled from public astronomical datasets and open science archives. Every object is rendered live in your browser using STEN, the Star Engine, a custom WebGL renderer that produces spectral-class-accurate stars with animated convection cells, coronal glow, and type-specific visual behavior.
The formal designation is Stellar Texture & Environment Navigator a name that describes the system's actual scope. It became STEN, and then informally the Star Engine. STEN is a real-time stellar environment simulation system: a physically parametric model of how light behaves in, around, and between stellar bodies. It characterises the spectral emission profile of a photosphere, the temperature gradient and opacity structure of an accretion disk, the distortion of background light by a compact object's gravity well, and the phase-dependent illumination of an orbiting planet all within a unified simulation. Every object in the catalog carries a physical description; STEN maps that description to a visual output. What you see is a direct consequence of spectral class, surface gravity, effective temperature and stellar mass, which is a real approximation, and not an artist's interpretation.
M-type Red Dwarf · STEN Render
Each object type has a dedicated shader pipeline from the blue-white fury of O-class supergiants to the silent darkness of stellar black holes.
Applications like Universe Sandbox offer a physics playground, and a genuinely excellent one. Collide worlds, sculpt planetary rings, watch stellar gravity strip matter from a companion across a binary orbit with excellent tools and options. We are not competing with that, and this was never the goal. STEN is a Stellar Object Renderer. Its purpose is to show you what a given star actually looks like: its spectral emission profile, its convective granulation, its limb darkening and coronal structure, all derived from its measured physical parameters. If Universe Sandbox is a flight simulator, STEN is precision satellite imagery of the real object. The Gaia catalog is the source seed data. STEN is what makes the picture. For a full feature comparison across TheStarDB, Stellarium, SkySafari and Universe Sandbox, see the comparison table on the STEN page.
The professional databases were built for researchers querying specific, well-studied objects. They are both invaluable and authoritative, and also extremely helpful but incomplete by design. Open any of the millions of entries in SIMBAD that lacks spectroscopic observation, and the spectral type field returns a single word: “Star”. HyperLEDA is another excellent dataset, and it catalogs more than four million galaxies; however it only contains morphological T-types for roughly 60,000 of them, excluding the greater part of 99.94% of its catalog. NASA/IPAC NED indexes approximately three billion source entries, but the overwhelming majority are simply typed Galaxy, with no morphology, stellar mass, or star formation rate. Consumer apps do not improve on this either, as no one has done the work. SkySafari, Stellarium, and even Universe Sandbox displays essentially no classification data for faint objects beyond a generic label. Nobody has built comprehensive photometric classification across 16M+ Stars, 17M+ Stellar Objects and 22M+ galaxies with computed physical properties for all of them, that is…until now.
Gaia DR3 publishes precision astrometry and broadband photometry for 1.8 billion sources. Utilizing the BP-RP color index, the flux ratio between Gaia’s blue and red passbands, we derived an effective temperature for all 16 million+ stars in the catalog using calibrated color-temperature relations from the referred literature. Temperature plus luminosity (absolute magnitude from parallax distance, corrected for bolometric flux) yields stellar radius via the Stefan-Boltzmann relation. From this derived temperature, we assigned a photometric spectral type, O through M with subtype, to every star in the catalog. All of this is standard stellar astrophysics. The incumbents had the same source data. They simply did not apply it at this scale for a public database. But we did.
For the 3,153 star clusters in our catalog, the same derivation principle was extended further. The Kharchenko+2013 MWSC catalog provides distances, ages, and member counts for 3,006 open clusters, yet integrated apparent magnitudesthe brightness an observer on Earth would actually measurewere absent for 85% of them. Rather than leave those fields blank, we applied a physically motivated empirical calibration: the integrated absolute magnitude of a cluster is predicted by MV = a + b·log10(N) + c·log10(t), where N is the member count and t is the age in years. Physics constrains the coefficient signs before any regression is run: b must be negative (more members contribute more total light, brightening the cluster), and c must be positive (older clusters have lost their most luminous O- and B-type stars to stellar evolution, dimming them). Fitting this model on the 447 clusters with confirmed catalog magnitudes yields MV = −8.72 − 1.04·log10(N) + 0.98·log10(t), with both coefficients obeying those physical constraints. A distance modulus conversionμ = 5·log10(dpc) − 5translates each derived absolute magnitude to the apparent brightness seen from Earth. Applied to 2,559 clusters without catalog values, this raises apparent magnitude coverage to 99.97% of the open cluster dataset, with a calibration residual of ±1.65 mag. The source data had existed in public catalogs for decades. The derivation is textbook astrophysics. No public cluster database had applied it at this scale before.
That being said, is a photometric classification of a Star or Galaxy as definitive as a measured spectrum? No. That is exactly why every entry is labeled: SPECTROSCOPIC where a real observed spectrum exists (roughly 119,000 named and nearby stars from HYG and other cross-references), and PHOTOMETRIC everywhere else. No other public astronomical database (including NASA) surfaces this distinction at the record level. The key is being honest about what the classification is, and we are, and always will be.
GLEN is on its way…
For galaxies and our upcoming GLEN program, we cross-matched five public catalogs: SDSS DR17 (1.1M spectroscopic redshifts and velocity dispersions), HyperLEDA (52k morphological T-types), HECATE (51k stellar masses and star formation rates), PGC2003, and Mangrove. Every galaxy in the resulting 22.4-million-entry dataset carries whatever the union of those five sources can provide. The photometric color-morphology relation, i.e., redder integrated color implies early-type tendency, and bluer shift implies late-type spiral, fills in estimates for the remainder of the data, and while this gives us an almost complete spec dataset for all 22M+ galaxies, we clearly and honestly flag this data as derived for later verification. As it stands, this is the first publicly queryable galaxy catalog at this scale with computed physical properties across the entire dataset. No one else has this, or even comes close.
STEN's geodesic raymarcher simulates real relativistic light bending around compact objects photon sphere, gravitational lensing, frame dragging, and a physically-modelled accretion disk computed entirely on the GPU, per frame, at real-time frame rates. No pre-rendered frames. No video playback. Pure WebGL2.
The free explorer gives you G-type stars, red dwarfs, white dwarfs, and pulsars a taste of what STEN can do. A full license opens every stellar class, every compact object, every exotic phenomenon the universe has to offer, rendered live in your browser at real-time frame rates.
No credit card required for free account · Full license unlocks all 16 stellar object types in STEN
No application to download. No driver to install. No sandbox to configure. No account required to explore. STEN is a full stellar renderer built entirely in WebGL2, open a tab, and the stars are already burning.
Chrome, Safari, Firefox, Edge. Desktop, laptop, Chromebook, tablet. If it has a GPU and a URL bar, it runs STEN. No GPU drivers to configure, no Vulkan, no DirectX. The browser handles all of it automatically.
Full touchscreen navigation built in from the ground up. Zoom, orbit, and select objects with your fingertips. Plug in any controller, Xbox, PlayStation, or generic USB, and navigate the stellar catalog with your thumbstick. No configuration. No drivers. It just works.
Every stellar class has its own ambient audio signature, white noise for quiet dwarfs, deep magnetic hum for magnetars, crackling plasma for red supergiants. Original ambient music plays throughout. All streamed from the browser. All optimised for looping.
A Raspberry Pi or a $60 refurbished mini PC. A cheap touchscreen. A browser pointed at thestardb.org. That is the entire setup for a full astronomy kiosk your students can explore with their hands. No expensive software licenses. No classroom lab subscriptions. No IT department required. STEN runs on hardware that schools already own.
STEN renders physically-modeled stars in real time using a handful of kilobytes of GLSL shader code. Convective cells, limb darkening, chromatic atmospheres, magnetic flares, all computed per-frame on the GPU. No gigabyte asset packs. No streaming textures. The entire renderer loads in seconds.
The TheStarDB API gives developers programmatic access to the full stellar catalog, spectral types, luminosities, distances, proper motion, and more. Build planetarium apps, educational tools, AR overlays, or research pipelines on top of real, structured stellar data. REST-first, JSON everywhere, rate limits that scale with your plan.
STEN (The Star Engine) is available for commercial licensing. Embed the renderer in your educational software, science museum installation, streaming platform, or custom kiosk application. STEN renders stellar objects in real time inside any modern browser, no plugin, no runtime, no backend GPU.
Browse the full stellar catalog, or jump directly into the STEN explorer and render any star in real time.