Space-Track Space Catalog API

Every artificial object humanity has put into orbit, held in one ledger. Space-Track is the public face of the US Space Force's 18th Space Defense Squadron catalogue and the primary authoritative source for space situational awareness data - built to promote space flight safety, protection of the space environment and peaceful use of space by sharing SSA services with satellite owner/operators, academia and other entities. The catalogue covers roughly 51,500 tracked objects: about 12,000 active payloads, some 22,900 debris objects, and around 16,600 analyst objects whose element sets are withheld from publication entirely.

Two things make it the reference rather than merely a source. First, depth: the historical archive holds more than 138 million element sets, so an object's orbital behaviour can be reconstructed across decades, not just observed today. Second, breadth beyond positions - the same system carries catalog-debut records for newly launched objects, decay events when something leaves orbit, reentry predictions, and conjunction data messages flagging pairs of objects heading toward each other. That combination turns a list of orbits into a research corpus. Get a sample of this dataset to inspect real rows before you commit pipeline time.

The row shape a delivery lands in - one row per tracked object, catalog identity joined to its latest element set:

# current catalog - one row per tracked object (row shape shown; every value pinned to real records in your sample)
norad_cat_id=25544   object_name=ISS (ZARYA)                    segment=active payload      epoch=<latest element set>
norad_cat_id=#####   object_name=<catalog designation>          segment=debris              decay_date=null
norad_cat_id=A####   object_name=<withheld designation>         segment=analyst object      element_sets=not published

# element-set columns riding along with each state vector
epoch=<datetime>  inclination=<degrees>  raan=<degrees>  eccentricity=<decimal>  argument_of_perigee=<degrees>
mean_motion=<revs/day>  mean_anomaly=<degrees>  bstar=<drag term>

# lifecycle and risk tables beside the live catalog
table=DECAY   norad_cat_id=#####   decay_date=<date object left orbit>
table=CDM     cdm_id=<message id>  tca=<close-approach time>

The three segments in the first block are the catalogue's own accounting, not our labels: roughly 12,000 active payloads, some 22,900 debris pieces, and about 16,600 analyst objects whose orbital elements simply do not circulate. That third row matters more than it looks - any vendor selling you "all 51,500" without flagging it is overselling. The middle block is the classical element set: six numbers plus a drag term that place an object in space and time, with mean_motion alone sorting the population into low, medium and geosynchronous regimes. And the bottom block is why operators treat this as infrastructure rather than trivia - decay history and conjunction warnings ship beside the live catalog, so the pull that describes an orbit also warns about the one bearing down on it.

Your sample pins every value above against real records before anything reaches your pipeline.

Ten core fields define every row: two identity columns (norad_cat_id, object_name) and the eight quantities of the classical element set, from epoch through mean_anomaly plus the bstar drag term. Definitions follow the published element-set convention, not guesswork. The deeper fields flagged underneath the table - creation timestamps, decay dates, registry attributes, conjunction identifiers and full history pulls - fold into your sample on request.

• Geography: global by construction - the catalogue tracks Earth-orbiting and deep-space objects from every launching state, and the per-country summaries let you slice the population by registry without any extra join.
• Temporal: the current catalogue sits on top of an archive of 138M+ element sets stretching back to the start of the space age, so trend work - how debris populations grew, when particular shells filled up - runs off the same source as today's state vectors. Decay and conjunction records carry their own retention windows.
• Granularity: one record per space object, with per-element-set history beneath it. An object aloft for decades has thousands of observations; a launch-week payload has a handful. The unit of analysis is the object-epoch pair.

Scale check: ~51,500 objects sounds modest next to tabular business data, but history multiplies it past 138 million rows - large enough that bulk archives arrive partitioned, small enough that the current catalogue loads whole.

API, files, or your warehouse. Daily, weekly, or hourly.

You choose the channel and the cadence; the field dictionary travels unchanged across all three. Payloads suit products that surface a single object's latest state inside an app, files suit analysts loading the full current catalogue once per study, and warehouse delivery suits teams joining orbital data against their own fleet registers in SQL. Bulk history lands pre-partitioned, so a decades-long reconstruction never becomes a pagination project. Changing cadence later is a settings conversation, not a re-integration.

• Conjunction screening support - conjunction data messages and per-object histories feed collision-avoidance workflows, letting teams watch the objects they care about without standing up their own SSA pipeline. One hard line applies everywhere: these mean elements must never be used for conjunction assessment prediction itself.
• Debris-environment research - the split between active payloads, debris and withheld analyst objects, traced through the historical archive, is the baseline series for anyone modelling orbital crowding or reentry flux.
• Fleet and constellation monitoring - inclination and mean motion separate regimes cleanly, so operators and insurers can watch competitors' shells, confirm deployments and spot deorbit campaigns from the outside.
• Registry and market analysis - per-country object counts and launch-site references turn the catalogue into a national space-activity panel that updates with every launch.
• Reentry and lifecycle tracking - decay events paired with reentry predictions cover the end-of-life tail most datasets stop at.

Data scientists and ML engineers get a keyed, per-object corpus with decades of history - ideal for propagation studies, anomaly detection and debris-growth forecasting. Developers and data-product builders get uniform element-set fields, so one code path covers every object ever cataloged. Investors and quants get constellation deployment and retirement visible as data rather than press releases. Journalists and academics get citable primary records under government custody instead of third-party mirrors. See the data scientists use cases page for workflow patterns and the developers builders use cases page for product-side detail.

Three things worth knowing upfront. First, these are general perturbations mean element sets - fitted values designed for the standard propagator, not raw precision ephemerides, so precision claims should be framed accordingly and they must never be used for conjunction assessment prediction. Second, the headline count includes about 16,600 analyst objects whose element sets are never published, so the effective catalogue available to anyone is smaller than the ~51,500 total - we flag those gaps explicitly rather than padding extracts. Third, the historical archive ships partitioned by year; if your use case needs only recent epochs, say so and the sample gets scoped accordingly.

Notes that pair well with this page:

• Launch Library 2 API (Space Devs) - launch events and mission context explaining each new arrival in the catalog.
• NASA Open APIs (api.nasa.gov) - mission feeds and earth observation complementing the orbital-state view.
• UNOOSA Register of Objects Launched into Outer Space - the treaty-based registry to reconcile against radar-tracked reality.
• SatBeams Satellite Coverage & Channel Database - commercial GEO detail - beams, transponders, footprints - layered over the same orbital backbone.
• Best aerospace & defense datasets - where this dataset ranks among the strongest sources for the sector.