compare_light_pollution
ChatGPTCompare 1–5 evaluated points. Default DarkSkySites API only. Browser transports must be explicitly requested and admitted; unavailable API never triggers browser fallback. Selected places require current revisions. Results retain dataset, coordinate binding and coverage uncertainty; this does not establish tonight’s weather or site access.
compare_observing_options
ChatGPTCompare two to five user-selected place/date/activity options. Each option uses one useful forecast provider by default; forecastMode compare or all represents an explicit request for independent or every source this tool can display. OpenWeather values require the visual planning card and are excluded here, including in all mode. Results keep sources separate, retain options with missing data, and show activity-specific factors without inventing a winner or universal sky score.
discover_observing_targets
ChatGPTOptional catalogue shortlist for a selected place and session time, independent of a weather forecast. Requires one resolved place snapshot, explicit shortlistLimit 1..6, localNightDate, and top-level displayTimeZone. A strict window accepts only UTC startUtc/endUtc, not window.timeZone. Optional failures remain explicit. Geometric source calculations use rounded coordinates; no terrain, weather or detection guarantee.
get_astronomy_window
ChatGPTCalculate astronomical darkness coverage of the requested window, nested Sun-below-threshold intervals clipped to that window, and a Moon phase/altitude summary for a resolved named place or revision-pinned candidate with explicit display time zone. This does not classify whole-day polar conditions. Terrain and horizon obstructions are not modeled.
get_aurora_outlook
ChatGPTRetrieve NOAA SWPC's short-horizon OVATION location/intensity forecast for a previously resolved place. The nearest coarse grid cell and forecast times are labeled. This is not a personal sighting probability and does not evaluate darkness, clouds, horizon, or local conditions.
get_catalogue_target_windows
ChatGPTGet geometric windows for one to six exact returned pinned starter referenceId/revision pairs. Requires the selected place snapshot, top-level localNightDate and displayTimeZone, and a strict window with only UTC startUtc/endUtc. Optional failures remain explicit. These source calculations are separate from the main local planner; no arbitrary targets or local engine catalogue expansion.
get_clear_sky_climatology
ChatGPTRead optional approved historical cloud evidence for the selected API snapshot when configured. Without approved data, historical context remains unavailable; no acquisition or download is triggered. This is not a forecast or whole-night success probability. Results include method, period, sample, and coverage limits.
get_location_context
ChatGPTReturn source-linked place context and explicit unknown states for unverified darkness, public access, night opening and safety.
get_night_conditions
ChatGPTRead point forecasts with server-formatted local/UTC cloud rows for focused sample inspection. Default forecastMode single queries one useful source, with at most one fallback if that source has no usable sample. Compare and all modes represent explicit requests for independent forecasts or every source this tool can display. OpenWeather values require the visual planning card and are excluded here, including in all mode. The selected named-place or revision-pinned candidate point is required; candidate planning also requires an explicit display time zone. Results keep source fields and units separate, identify missing variables, and include forecast steps overlapping astronomical darkness plus Moon illumination. Forecasts are not observations and may differ from local conditions.
get_observing_location_details
ChatGPTInspect a selected immutable API source snapshot by placeId and sourceRevision. Expired or missing snapshots require new resolution/search. Source coordinates are retained exactly; unknown site timezones require an explicit display zone for calculations.
get_reference_details
ChatGPTRead exact selected objects, meteor context, Earth close approaches or exoplanet hosts. Maximum two requested sections. Meteor context is not an annual forecast; exoplanets are learning context.
get_target_visibility
ChatGPTCalculate geometric above-horizon intervals and up to three timestamped orientation samples for the Sun, Moon, planets, or Milky Way core at a resolved named place or revision-pinned candidate with an explicit display time zone. The window requires endUtc after startUtc. Returned samples include IANA local labels tied to their exact instants. Existing samples at the same place, date, and instant provide the same evidence. Terrain, obstructions, atmosphere, and equipment limits are not modeled.
plan_observing_session
ChatGPTCombine a selected city or revision-pinned candidate point and display time zone, astronomical darkness, Moon conditions, sky events, a quota-saving forecast selection, and activity-specific guidance. Default forecastMode single uses one useful provider with at most one fallback; compare represents an explicit independent-forecast comparison and all represents a request for every forecast this plan can display. OpenWeather samples are available only when the visual planning card is enabled. A resolved place is required; an unverified candidate site zone also requires a user-chosen display time zone. Results keep source evidence and unknowns separate; forecast conditions are not guarantees, and access, seeing, transparency, and site safety are not inferred.
resolve_observing_region
ChatGPTResolve named countries or states into Geoapify boundary-filter identities. Results include regionId and sourceRevision for scope.regionSelections; ambiguous matches require a selected result. Region centers are not observing points. Unsupported memberships and exclusions remain explicit.
resolve_place
ChatGPTResolve a named place using the configured bounded GeoNames/Geoapify APIs. Results include sourceRevision and freshness. Ambiguous matches require a selected result; an unknown actual-site time zone requires an explicit display zone for later calculations. Coordinates and street addresses are not accepted.
search_observing_locations
ChatGPTSearch bounded Geoapify POI records using selected API origins and sourceRevision, or selected country/state provider boundaries with regionSelections. Return source snapshots and distance bands; unsupported exclusions and administrative intersections remain explicit. The source does not verify night access, safety or darkness.
search_reference_objects
ChatGPTLexical lookup of one exact object name or catalogue ID per query in one domain, at most ten matches. It does not perform semantic free-text, multi-object OR, or observing shortlists. Results provide opaque references for selected details. Catalogue identity does not imply local engine support or visibility.
search_sky_events
ChatGPTFind calculated Moon phases, solar/lunar eclipses, equinoxes, solstices, outer-planet oppositions, explicitly defined inferior/superior solar alignments, Mercury/Venus greatest geocentric three-dimensional Sun-angle elongations, and numerically screened geocentric right-ascension planet-pair conjunctions for all 21 Mercury–Neptune pairs within 2025–2027, plus sourced year-specific meteor forecast maxima. Pair coverage is partial: exhaustive completeness and stationary/tangent contacts are not certified; unresolved intervals are reported. RA crossings, solar alignments, elongations and closest approaches are distinct. Planet local visibility, darkness and detectability are not established; solar events do not authorize unfiltered solar viewing. Lunar intervals carry model-conditional horizon bounds; solar contacts use the stated refraction model. Separate source-backed verbal calendar guidance describes typical annual periods, not exact dates, requested-window activity or radiant visibility. Meteor season and radiant coverage are limited. Unsupported classes and coverage gaps are explicit.