analyze_relative_motion
ChatGPTPlot separation distance and radial/in-track/cross-track (RIC) components between satellites in ANY orbit class, including LEO and GEO. Use this for 'distance between HAWK-A and HAWK-B over the past 24 hours', relative motion, or how far apart satellites are over time. Resolve names to unambiguous NORAD IDs with satellite search, then pass nested request with kind=separation, primary and secondary catalog sources, start_iso, end_iso and step_seconds (default 60). For several comparisons use secondaries instead of secondary. Returns exactly two time-series charts: separation magnitude in km and signed R/I/C in km. Do not substitute GEO longitude/latitude history or a text sparkline. Windows are at most 72 hours; closest separation is a sampled minimum. Catalog sources use the latest TLE at or before the window start. This profile supports separation only, not conjunction screening.
analyze_visibility
ChatGPTAnalyze when satellites are visible from one or more ground locations: geometric passes above an elevation mask, and the intervals when the site is dark while the satellite is sunlit, so an observer could see it. Optional Open-Meteo weather filters those optical intervals. Pass a nested request with one source (kind catalog with NORAD IDs, selector for a catalog filter, or tle for caller TLEs), locations as explicit coordinates or place names, and one to ten windows. caps allow at most 25 satellites, 10 locations, 20,000 sample points and 60 seconds here; windows default to 24 hours and may be raised to seven days with max_window_seconds. Resolve satellite names with search first. Never infer the user's location: ask for coordinates or a place, and confirm ambiguous places. Runs immediately and creates no job. Results are TLE-based predictions; they do not assess imaging, SAR or RF access, or operational safety.
build_coorbital_neighbor_table
ChatGPTBuild a bounded co-orbital/co-planar neighbor table for one reference satellite or TLE and optional explicit candidate NORAD IDs. The nested request.reference selects the reference by NORAD ID, query, or explicit ID set, while request.candidates optionally restricts the candidate set. Use this for catalog neighbor discovery; use analyze_relative_motion for caller-supplied pair geometry. Typical requests include finding co-orbital neighbors of the ISS or comparing two explicit satellite IDs. Use similarity_mode=coplanar to match orbital planes, or coorbital to also constrain semi-major axis and eccentricity. Supports every orbit class.
build_geo_history_plot
ChatGPTPlot GEO satellite latitude/longitude over a requested UTC interval using successive historical TLEs propagated onto one common time grid. Supply satellites for explicit GEO-like tracks, or reference (for example Luch-2) for automatic neighbors. Non-GEO selections are rejected. For separation distance or RIC plots between satellites in any orbit class, use analyze_relative_motion with kind=separation instead. This tool does not return a separation/RIC time series. Automatic discovery propagates the caller-visible GEO-like catalog and retains objects meeting longitude and optional true 3-D distance gates at any common sample; current longitude is not a prefilter. Source epochs, TLE handoffs, stale/missing samples, candidate coverage and omitted matches remain explicit. Encounters between samples can be missed. Use this for longitude-over-time waterfall/history requests; use build_geo_neighbor_table for a current catalog neighbor snapshot.
get_satellite_tles
ChatGPTRetrieve complete TLE line1/line2 and source epoch for up to 100 selected NORAD IDs in request order. Resolve names/cohorts through search_satellites first. With as_of_iso, return the latest available historical TLE at or before the time; never substitute a future TLE. Preserve unavailable identities and partial results. Latest available does not guarantee freshness; report the actual source epochs.
list_satellite_facets
ChatGPTInspect ALL observed values through pages, including missing metadata, before translating synonyms into catalog filters. Countries include readable labels and affiliation codes; businesses include names, UUIDs, owner codes, government flags and associated satellite catalog countries. Counts are distinct NORAD IDs within the current filters. Business countries are catalog associations, not nationality. Keep the same request and use next_offset until enough vocabulary is known. Military/reconnaissance classifications require catalog evidence; government is not a synonym for military. Exact country matching is the default; affiliation mode explicitly includes legacy/shared source codes. No raw SQL is needed.
satcat_capabilities
ChatGPTAnswer 'What can SATCAT do?', help and example-prompt questions for this ChatGPT app. Recommend only available workflows and use the returned natural-language examples. Ask for missing dates or observer coordinates. Do not advertise broader SATCAT website features or unimplemented plots. Also supplies data-source notices when the user asks about attribution or sources. No catalog query, computation or write is performed; no visualization is opened.
satcat_compare_satellites
ChatGPTCompare current catalog orbital elements for two to four selected NORAD IDs. Epochs can differ; display each source epoch. Resolve names with satellite search first.
satcat_get_profile
ChatGPTReturn the SATCAT profile represented by this request's authenticated credentials. The opaque id is unique within this app and remains unchanged across token refresh, reconnection, and display-metadata changes.
satcat_globe_catalog_page
ChatGPTLoad the next immutable caller-owned globe page; never execute a new search.
satcat_globe_filter_options
ChatGPTRead current caller-visible categorical vocabulary for globe filters, including country, mission, object, orbit and operational status. Use exact returned values when translating military, constellation or orbital qualifiers. For business names use filters.business_name; it includes company subsidiaries. Never assert a classification beyond the catalog labels.
satcat_ground_passes
ChatGPTCompute geometric satellite passes above a specified elevation mask for explicit observer coordinates and a time window of at most seven days. Uses the latest catalog TLE at or before the window start. Ask for missing observer coordinates; never infer the user's location. Does not assess optical visibility, weather, or operational safety.
satcat_orbit_history
ChatGPTRead catalog TLE orbital-element history over an explicit UTC/offset window of at most 365 days. Returns at most 1,000 samples and reports truncation. No propagation or maneuver inference. Pass every requested element in metrics (e.g. ["apogeeAltitudeKm", "inclinationDeg"]) to preselect them together in one plot. The user can adjust the multi-select dropdown in the iframe.
satcat_search_satellites
ChatGPTFind satellites by name, NORAD ID, international designator, or owner. Return a paginated table; the iframe fetches subsequent pages on demand. For complete cohorts (e.g. Russian military), supply typed country/owner/business/mission filters, using source-compatible values; query is then the display label. Supply an identity search or at least one restrictive filter; empty filters and selecting both orbit statuses alone are rejected. Total matches are uncapped. Do not enumerate every NORAD ID or treat the first page as the whole catalog. Use selected NORAD IDs in other SATCAT tools.
satcat_separation_series_page
ChatGPTLoad one page of a caller-owned separation series for the analysis viewer.
satcat_show_globe
ChatGPTShow satellites on the Afterburner Cesium globe using satellite.js WASM SGP4. Select all_catalog=true for every caller-visible on-orbit satellite, filters for a cohort (including business ownership and orbital bounds), or mixed NORAD IDs, international designators and names in identifiers. Numeric IDs stay exact; ambiguous names require user clarification. Qualifiers combine with AND; values within a facet combine with OR. Use satcat_globe_filter_options for catalog values. No 25-result search cap: the widget pages the entire captured selection (resource ceiling 100,000 source TLEs, reported as an error if exceeded). Without a date, use current TLEs. For 'during DATE', provide start_iso: choose the closest historical TLE to that UTC date, earlier source on an equal-distance tie. For overlapped historical TLEs, provide history_start_iso/history_end_iso and explicit satellite identities. If missing, FIRST ASK the user to define the common propagation date; never infer it from the source history range. Supply that date as start_iso. History overlays allow 2,000 TLEs; narrow the window if exceeded. Filters use current catalog metadata. Show source epochs and warnings. Display positions are not authoritative visibility, conjunction or operational safety analysis. Read-only.
search_satellites
ChatGPTResolve satellites by any combination of country, business, owner, mission, object/status/orbit type, site, manufacturer, bus, NORAD ID or name/designator. Use list_satellite_facets to inspect actual values and map user synonyms yourself. AND across facets, OR within each; selected businesses include subsidiaries by default. Return NORAD/RSO IDs, directory identities and TLE availability/epochs. Empty filters browse all caller-visible objects; no implicit payload, operational or on-orbit filter. total_count describes the whole selection, not the first page. Follow next_offset; retrieve chosen TLEs with get_satellite_tles.