rf-los-mcp
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@rf-los-mcp Can I hit repeater W6ABC at 37.8044,-122.2712 from 37.7749,-122.4194 on 146.52 MHz?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
rf-los-mcp
A FastMCP (stdio) server that computes RF line of sight over locally cached
FABDEM V1-2
terrain (30 m bare-earth elevation). It is meant to be combined with a
repeaterbook MCP server: take repeater lat/lon and output frequency (MHz) from
repeaterbook, then ask this server whether you can reach the repeater from a
point or along a route.
Setup
python3 -m venv .venv && .venv/bin/pip install -e .Register with Claude Code (run this yourself; it is not done by the install):
claude mcp add rf-los -- /absolute/path/to/.venv/bin/rf-los-mcpRelated MCP server: mcp-opentopodata
Tools
Tool | Purpose |
| LOS between observer A and target B, with Fresnel clearance when a frequency is given |
| Coverage of one or more targets along a route (waypoints, a GPX file or a GeoJSON file) |
| Ground elevation (m AMSL) for up to 1000 points |
| Elevation profile along a path (max 20,000 samples), optional earth-curvature adjustment |
| List cached terrain tiles and the cache directory |
| Download tiles for a bounding box or centre+radius so you can work offline |
Defaults: observer 1.5 m AGL, target 25 m AGL (heights are AGL unless
*_height_ref="amsl"), k-factor 4/3, "clear" means 0.6 x first Fresnel zone is
unobstructed, LOS sample spacing 30 m, route spacing 200 m, target snap radius
1000 m, route max range 100 km. Units are metres, kilometres (route distances
only), MHz and decimal degrees.
There is no default frequency: without frequency_mhz the check is
geometric only and the result says so. Snapping moves the targets (repeater
coordinates are often imprecise) to nearby higher ground, never the observer.
For out-and-back routes pass the outbound leg only. Route waypoints and GPX or
GeoJSON points are geometry, resampled every sample_spacing_m along the route; only
named waypoints plus the start and end get per-waypoint verdicts, so name
the stops you care about (GPX track points are treated as unnamed). A route
call is limited to 50 targets, 10,000 line-of-sight checks (samples x targets)
and about 5 million terrain samples in total, including target snapping.
Other limits: rf_los_check paths and rf_los_route's max_range_km are at
most 300 km. Any one call (other than dem_prefetch) downloads at most 16 new
terrain tiles (about 200 MB), and dem_prefetch takes at most 100 tiles
(about 1.2 GB) per call. These limits stop a mistaken or manipulated agent
request from filling the disk or hammering the FABDEM server.
geojson_path takes a .geojson/.json file (for example an OpenStreetMap
export). It must pass the official
GeoJSON JSON Schema (RFC 7946,
bundled; MIT licence), so any other JSON file is rejected before its contents
are used. Prefer it to copying GeoJSON coordinates into waypoints, because
GeoJSON stores positions as [lon, lat]. LineString and MultiLineString pieces
are joined end to end (pieces may be reversed or out of order, but branching or
disconnected pieces are rejected). Point features with a name property, such
as a summit or hut, are placed on the line as named waypoints and must be within
500 m of it. A file with only Points uses them, in order, as the route.
Caveats
Vegetation and buildings are not modelled (bare-earth DEM), so results may be optimistic through dense forest.
Route analysis samples the route and the line of sight at discrete spacing (default 200 m along the route); narrow ridges between samples can be missed.
Ocean is treated as sea level (0 m). A missing land tile is an error, never 0.
Terrain cache
Tiles are cached in $RF_LOS_CACHE_DIR, else $XDG_CACHE_HOME/rf-los, else
~/.cache/rf-los. Each 1 degree square tile is about 12 MB. The first query in
a new area downloads the tile(s) it needs, which can take about a minute per
tile. To avoid that (or to work offline), prefetch before you go, e.g. ask the
agent to call dem_prefetch with a bounding box (south, west, north,
east) or center_lat, center_lon, radius_km; areas needing more than
max_tiles (default 16) tiles are refused. Offline queries over cached tiles
work normally; uncached land tiles produce an error naming the tile.
Up to 6 decoded tiles are held in memory (about 52 MB each, ~310 MB peak).
Attribution and licence
Code: copyright (C) 2026 Aaron Brown, licensed under the GNU Affero General Public License v3.0 only (AGPL-3.0-only); see LICENSE. If you run a modified version as a network service (for example a hosted MCP server), you must offer its users the source of your version under the same licence.
GeoJSON schema: src/rf_los/data/GeoJSON.schema.json is from
geojson/schema, copyright (c) 2018 Tim
Schaub, MIT licence (see src/rf_los/data/GeoJSON.schema.LICENSE.md).
Terrain: FABDEM V1-2 (Neal & Hawker 2023, University of Bristol, doi:10.5523/bris.s5hqmjcdj8yo2ibzi9b4ew3sn), licensed CC BY-NC-SA 4.0, derived from Copernicus DEM GLO-30. FABDEM is for non-commercial use only; do not redistribute the downloaded tiles. Results carry this attribution.
Available Tools
6 toolsdem_cache_statusDem Cache StatusA
List the terrain tiles cached locally (name, bytes) and the cache directory.
Tiles are named by their south-west corner, e.g. S34E151 covers latitude -34 to -33 and longitude 151 to 152.
Inputs: no named parameters. Returns: tiles, cache_dir, tile_count, total_bytes
dem_prefetchDem PrefetchA
Download terrain tiles for an area so later queries work offline.
Give EITHER a bounding box (south, west, north, east; decimal degrees)
OR center_lat, center_lon and radius_km. Each 1 degree tile is about
12 MB and can take about a minute to download. If the area needs more
than max_tiles (default 16, at most 100) tiles, nothing is downloaded
and an error states the count; raise max_tiles deliberately. Tiles that fail to
download are listed in failed (retry later); ocean tiles need no data.
Inputs: east, west, north, south, +4 more. Returns: ocean, failed, requested, downloaded, +2 more
rf_los_checkRf Los CheckA
Check RF line of sight between observer A and target B over FABDEM terrain.
A is the observer (default 1.5 m above ground level); B is the target, e.g. a repeater (default 25.0 m AGL). Coordinates are decimal degrees, heights are metres. a_height_ref/b_height_ref are "agl" (above ground level, the default) or "amsl" (above mean sea level).
There is NO default frequency. Without frequency_mhz the check is geometric only (mode="geometric", no Fresnel zone) and the result says so. With frequency_mhz the 0.6 x first Fresnel zone must be clear for a "clear" verdict (mode="rf"); partial intrusion is "marginal". Take repeater lat/lon and output frequency (MHz) from the repeaterbook tools.
k_factor is the effective earth radius factor (default 4/3, standard atmosphere). snap_radius_m (default 1000 m, max 10000, 0 disables) snaps the TARGET B (not the observer A) to the highest nearby ground within that radius, since repeater coordinates are often imprecise (valley town vs the hilltop site); the move is reported in target_snap. A and B must be different points at most 300 km apart. Vegetation and buildings are not modelled (bare-earth DEM). The first query in a new area downloads a ~12 MB terrain tile per 1 degree square, which can take about a minute; use dem_prefetch to pre-download. One call downloads at most 16 new tiles.
Inputs: a_lat (required), a_lon (required), b_lat (required), b_lon (required), +7 more. Returns: mode, notes, summary, verdict, +15 more
rf_los_routeRf Los RouteA
Assess RF coverage of target repeaters along a route (e.g. a drive or hike).
Provide the route as EXACTLY ONE of waypoints (list of lat/lon, in
order), gpx_path (path to a .gpx file of at most 10 MB readable by
this server, ~ expanded; track points, else route points, else
waypoints) or geojson_path (path to a .geojson/.json file of at most
10 MB, e.g. an OpenStreetMap export; prefer this over copying GeoJSON
coordinates into waypoints, which risks swapping GeoJSON's [lon, lat]
order). The file must be valid RFC 7946 GeoJSON (checked against the
GeoJSON JSON Schema; other JSON is rejected). GeoJSON
LineString/MultiLineString pieces are joined end to end
(branching or disconnected pieces are an error); Point features with a
name property become named waypoints placed on the line (within
500 m); with no lines, the Points are the route. The route must have
non-zero length. Each target has lat, lon, optional name, height_m
(default 25.0 m AGL), height_ref ("agl" or "amsl"), and optional
per-target frequency_mhz / snap_radius_m which override the call-level
values.
Waypoints / GPX points are route GEOMETRY: the route is resampled every
sample_spacing_m metres (default 200) along its length, so dense GPS
tracks are fine. Per-waypoint verdicts are reported ONLY for NAMED
waypoints plus the first ("Start") and last ("End") point: give a
name to each stop you want a result for (summit, hut, camp). GPX
track points are unnamed geometry; GPX route points and waypoints keep
their names. Line of sight is evaluated at the samples, an
approximation that can miss narrow ridges between them. At most 10000
line-of-sight checks (samples x targets) per call; for long routes or
many targets increase sample_spacing_m.
observer_height_m is the antenna height AGL along the route (default 1.5 m). Route positions are in km. Targets farther than max_range_km (default 100, max 300) from a point are reported out_of_range.
There is NO default frequency: without frequency_mhz the result is geometric only (no Fresnel zone) and says so. Take repeater lat/lon and output frequency (MHz) from the repeaterbook tools. Snapping (snap_radius_m, default 1000 m, 0 disables) moves the TARGETS to higher nearby ground, never the observer waypoints.
For an out-and-back route pass the OUTBOUND LEG ONLY: the return leg repeats the same coverage. Output has per-target coverage segments with percent clear/marginal, verdicts at Start/End and named waypoints, and coverage gaps. Vegetation is not modelled. A first query in a new area downloads ~12 MB per 1 degree terrain tile (about a minute each; at most 16 new tiles per call); use dem_prefetch. At most 50 targets per call, and at most 5,000,000 terrain samples in total (target snapping included).
Inputs: targets (required), gpx_path, k_factor, waypoints, +6 more. Returns: gaps, notes, summary, targets, +7 more
terrain_elevationTerrain ElevationA
Ground elevation in metres above mean sea level (FABDEM, bare earth).
Up to 1000 points (decimal degrees). Sea and areas not in the dataset read as 0. Missing land tiles are downloaded on demand (~12 MB per 1 degree square, at most 16 new tiles per call) or raise an error naming the tile; see dem_prefetch.
Inputs: points (required). Returns: points, attribution
terrain_profileTerrain ProfileA
Elevation profile (metres AMSL) along a path through 2 or more points.
Samples every spacing_m metres (default 30) along great-circle segments between the points; every input point is a sample. Returns distance in metres from the start, lat/lon and elevation per sample. With include_curvature=true each sample also has curvature_adjusted_m = elevation + d1d2/(2k*R), the terrain as seen against the straight first-to-last chord, with earth bulge for k_factor (default 4/3). At most 20000 samples; for longer paths increase spacing_m. At most 16 new terrain tiles are downloaded per call.
Inputs: points (required), k_factor, spacing_m, include_curvature. Returns: points, attribution, sample_count, total_distance_m
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 6 tools
Each tool targets a clearly distinct operation: point-to-point LOS (rf_los_check) vs route coverage (rf_los_route), point elevations (terrain_elevation) vs path profiles (terrain_profile), and cache inspection (dem_cache_status) vs tile download (dem_prefetch). The descriptions explicitly cross-reference each other (e.g. 'use dem_prefetch to pre-download', 'prefer this over copying GeoJSON'), which further reduces misselection risk.
All names are snake_case and grouped by domain prefix (rf_los_*, terrain_*, dem_*), which is highly predictable. The only minor inconsistency is that most names are noun-phrase labels while dem_prefetch leads with a verb, but the overall pattern is still readable and coherent.
Six tools is well-scoped for an RF line-of-sight plus DEM server: two analysis tools, two terrain-data tools, and two cache-management tools. Every tool earns its place, with no redundant or filler entries.
The surface covers the core lifecycle: terrain data acquisition (dem_prefetch), inspection (dem_cache_status), elevation queries, profiles, and both point-to-point and route LOS analysis. Minor gaps exist, such as no fan-out check of multiple targets from a single fixed observer, no way to clear/purge cached tiles, and no direct link-budget or best-repeater recommendation, but agents can work around these by looping calls.
Maintenance
Related MCP Connectors
Look up elevation worldwide, profile route ascent/descent, grid areas, check terrain line of sight.
OpenStreetMap queries, maps/styles, search, routing, terrain, analysis, pipelines, and rendering.
Elevation, slope, terrain maps & 3-D surfaces for any site — near-global 30 m, US 10 m, AU/NZ 5 m.
Open-Elevation MCP — terrain elevation by lat/lon (no auth)
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables advanced geospatial analysis including elevation queries, line-of-sight calculations, terrain visualization, and autonomous drone mission planning with real-time safety analysis and 3D visualization capabilities.MIT
- AlicenseNot gradedqualityCmaintenanceProvides elevation data and DEM dataset listings via Open Topo Data, enabling querying elevation at points or listing available datasets.117 npmMIT
- AlicenseNot gradedqualityCmaintenanceEnables geocoding, reverse geocoding, elevation profiles, static map PNGs, and coordinate reprojection using MapTiler's OpenStreetMap data.85 npmMIT
- AlicenseAqualityBmaintenanceProvides 22 geospatial tools for coordinate conversion, geocoding, routing, terrain analysis, weather, and radio link planning, all without requiring API keys.22MIT