Skip to main content
The mission editor lets you author a route without connecting to a robot. You choose places on a map, inspect the planned route, and save it to GRID cloud. Then run that saved mission through GRID CLI.

​
1. Open the editor

Use the GRID CLI with your account signed in and the intended cluster selected:
Already in the GRID interactive shell? Run nav start. To print the URL without opening a browser automatically:
Open the exact http://127.0.0.1:PORT/navigation?session=... URL printed by the CLI. Keep that process running. The page uses the CLI’s existing login; you do not paste a GRID cloud API key into it. Each page is pinned to the account and cluster used when it opened. After switching either, open a new editor.
The mission editor authors maps and routes. The Visualization page observes a running robot. Saving or planning a mission in the editor does not drive a robot.

​
2. Choose or build a map

Select a COMPLETED map in Active map. Use Refresh to fetch changes and Load more maps when another page is available. A map still being built, or a failed build, cannot be selected for planning. If you have a scan from GRID Navigator, expand Build a map, enter a map name, select the iPhone scan ZIP, and choose Upload & build. The ZIP must be nonempty and at most 4 GiB. Keep the page open until upload and submission finish. The editor then polls build progress; closing the page after submission does not cancel the cloud build. Refresh the list and select the map once it is completed. The screenshot below uses an existing office map and its saved plant_door mission. Your map, names, and coordinates will differ. Mission editor showing a completed office map, the Start and WP1 waypoints, and a saved route.

​
Understand the map before placing points

In 2D · Binary, black is obstacle, white is observed free space, and grey is unknown. Use 2D · Color for captured color or 3D · Voxels to inspect the volume. Place waypoints in a 2D view. Drag empty space to pan, scroll to zoom, and use Fit map to recover the full map. Expand map hides the side panels. In 3D, drag to orbit, right-drag to pan, and scroll to zoom. 2D height slice selects the minimum and maximum heights used to build the occupancy view. Values are metres in the map frame, not distance above the floor; negative values can be correct. Enter a range within the scanned height extent and select Apply height. Pick a slice that captures the obstacles relevant to your robot. Changing it invalidates the route preview: check the waypoints and plan again. If a slice is empty, adjust its limits.

​
3. Place and name waypoints

Click observed free space. The Name this place dialog opens with an automatic name such as wp_1. Give it a useful unique name, inspect Map X and Map Y in metres, and choose Apply waypoint. The first point becomes the start. Later points become destinations. Add opens the same dialog when you want to type known coordinates rather than click. Enter verified coordinates: the initial numeric values are not a recommendation for where to navigate. Waypoint editor with map coordinates and an optional 90 degree heading. A heading is optional. Type degrees or drag the direction dial: 0° points along +X; 90° points along +Y. Blank means no heading requirement; zero is an explicit direction. Clear heading removes it. The current robot executor aligns to a heading at the final destination only. To adjust a waypoint, drag its marker, drag its heading handle, or use Edit for exact values. Names must be unique, at most 200 characters, and cannot contain angle brackets or a null character. Keep every waypoint on observed free space.

​
Save the place library

Save waypoints stores the named waypoint library on this computer, scoped to the CLI account, cluster, and map. Reload restores the last saved library and clears the current route. Unsaved changes trigger a warning; if another tab saved first, reload before editing again. This local save does not create a GRID cloud mission. Local names are not semantic labels searchable through nav.find(), and there is no remote nav.list_waypoints() method. Saved GRID cloud missions may label their stops WP1, WP2, and so on when loaded; local names remain in the local library.

​
4. Find a mapped object instead of choosing XY

Under Find an object, enter a description such as “door” and select Find. Review the returned matches and suggestions, then select the object you intend. This searches the stored map, not a live camera image. The orange marker is the object. With a start position selected, the editor can plan a reachable approach waypoint, shown in green, near that object. Inspect it in the waypoint dialog before applying it; Face object sets a heading toward the object. If no start is set, a free-space object can be added directly. Otherwise choose a start and use Generate approach waypoint. An object can be inside an obstacle or unreachable from the chosen start. An object’s location and the robot’s approach location are different things. Use Clear results when finished with the search.

​
5. Build and preview the route

In Plan a mission, enter a name and choose Start position. Use Use in mission beside an existing waypoint to add it to Destinations. The remove button removes a stop from this route; Delete in the waypoint library removes the waypoint itself. Choose Plan route. Review the line through free space, route length, destination count, and any green rings showing relocated destinations. Edits to waypoints, route settings, or the height slice require another plan.

​
6. Save the mission to GRID cloud

After planning, choose Save mission to Cortex. This is the save that makes a route available to nav.list_paths(map_id) and nav.run_mission(path_id). A plan preview alone is not a saved mission. Planned route with separate Save waypoints and Save mission to GRID cloud buttons. Check the success message and saved mission ID, then verify the mission appears under Saved missions. The editor checks saved heading metadata too. If it reports a saved mission with a verification warning, inspect the saved entry before saving again. A lost response or failed list refresh does not prove that the save failed; refresh the list first to avoid duplicates. Selecting an entry under Saved missions loads its route, height slice, and planning settings. It keeps the local library and adds any needed temporary points. Editing, replanning, and saving creates a new mission, rather than updating the original in place.

​
7. Run it from Python

Continue with Run a saved mission. Use the same map and the saved path ID. Place the physical robot at the mission’s intended start, start navigation, verify localization, and then run the mission. The editor does not perform these robot-side steps for you.