Updated October 2026 · Navigation & mobile mapping
AI can help you plan a trip, but it cannot replace an offline map. Conversational search, AI route builders and review summaries generally need a connection, while downloaded maps keep working when the signal disappears. This guide compares 18 navigation and travel apps, with particular attention to what still works in airplane mode.
What changed since 2024
Since our 2024 guide, AI has moved from a novelty to a mainstream feature in navigation and travel apps. Google Maps added Gemini-powered navigation and Ask Maps, Waze introduced Gemini-based reporting and a motorcycle mode, and AllTrails expanded its AI Custom Routes feature. Other apps, including Komoot and Strava, now use AI to help users discover or refine routes.
The key distinction is not whether an app has AI, but whether the AI output can be downloaded and used without a connection. A route created online is useful only if you can save it to an offline map before leaving.
Quick picker: which app fits your trip?
Tap a filter. The dots indicate how much of the app remains useful without a signal:
●●●●● Full offline navigation and search
●●●●○ Strong offline navigation, with some connected features unavailable
●●●○○ Limited offline use
●○○○○ Mostly online
Switch your phone offline
Flip the switch: green features work offline, purple features require a connection. “Offline” means the feature can be used after the necessary map, route or itinerary has been downloaded in advance.
| App | Best for | Features | Offline cost |
|---|---|---|---|
| Google Maps | Driving, cities | Downloaded areasTurn-by-turnOffline place informationLive trafficGemini in navigationAsk Maps | Free |
| Waze | Commuting | Route loaded before signal lossCrowd alertsGemini voice reportingRerouting | Free |
| HERE WeGo | Road trips abroad | Country / region / continent downloadsTurn-by-turnOffline navigationLive transitIncident reporting | Free |
| Apple Maps | iPhone users | Downloaded map areasTurn-by-turnPlace cardsNatural-language searchLive transit | Free |
| Wanderlog | Trip planning | Saved itinerary and map, where supportedAI assistantRoute optimisationLive collaboration | Pro subscription required for offline access |
| Citymapper | Urban transit | Saved journeysMetro maps, where availableNew route planningLive disruptions | Free |
| Tripadvisor | Choosing what to see | ReviewsAI trip builderAI review summaries | Not primarily an offline navigation app |
| MAPS.ME | Rural tourism | Country downloadsTurn-by-turnPOI searchContoursHotel booking | Freemium; check current in-app pricing |
| Organic Maps / CoMaps | Privacy, hiking | Country downloadsTurn-by-turnSearchOSM editing, synced later | Free |
| Komoot | Cycling, hiking | Downloaded region mapsVoice navigationClimb detection@komoot in ChatGPTRoute planner | Free region; additional regions or Premium |
| Strava | Training | Offline maps for subscribersGPS recordingAI route suggestionsAthlete Intelligence | Subscription |
| Ride with GPS | Long rides | Offline routesVoice cuesWaypoint alertsCollaborative planning | Paid tiers |
| REVER | Motorcycle touring | Downloaded routesRide trackingCommunity routesWeather overlay | Freemium / Pro |
| Scenic | Motorcycle day rides | Offline mapsNavigationRoute sharing | Freemium |
| AllTrails | Hiking | Offline trail maps, PlusWrong-turn alertsSaved custom routesAI Custom Routes builderConditions forecastOutdoor Lens | Plus or Peak |
| MapMyWalk | Fitness walking | GPS trackingBasemapChallenges | Freemium |
| OsmAnd | Off-grid power users | Country downloadsTurn-by-turnContours and hillshadeGPX tracksHourly map updates, Pro | Free core; paid Maps+ and Pro plans |
Based on vendor documentation and app-store listings checked in October 2026. Availability and prices vary by country, platform and subscription.
Drivers
Google Maps: The all-rounder UPDATED
- Offline: Download selected areas for turn-by-turn directions and offline place information. Live traffic, alternative routing and connected AI features are unavailable offline.
- AI: Gemini in navigation can help with stops and route-related questions while driving, walking or cycling. Ask Maps adds conversational search for places.
- Availability: Ask Maps began rolling out in the United States and India. Feature availability may differ by device, language and account.
Waze: The crowd-sourced navigator UPDATED
- Offline: Waze does not offer a full official offline-map mode. A route loaded before losing signal may continue from cache, but rerouting, search and crowd reports require a connection.
- New: Waze has added Gemini-powered conversational reporting and destination search, plus motorcycle mode in selected countries.
- Motorcycle mode: Google lists Argentina, Brazil, Colombia, Malaysia, Mexico, Peru and the Philippines among the initial markets.
HERE WeGo: The offline champion
- Offline: Download regions, countries or continents in advance for offline navigation.
- Online: Live transit information and incident reporting require a connection.
- Best for: Travellers who want a simple, free offline backup for driving and public transport.
Apple Maps: The iPhone default NEW IN GUIDE
- Offline: Download map areas for turn-by-turn driving, walking, cycling and transit directions, including on Apple Watch where supported.
- Online: Natural-language search, live transit and other connected features require a connection.
- Availability: iPhone only. Offline-map coverage and feature availability vary by region.
Travellers and trip planning
Wanderlog: The trip planner that puts everything on a map NEW PLAYER
- Best for: Building an itinerary around places, bookings and a map view.
- Offline: Offline access is part of the paid Pro plan. Confirm the exact offline behaviour for your itinerary and device before relying on it.
- Limitation: Wanderlog is a planner, not a full turn-by-turn navigator. Pair it with Google Maps, Apple Maps or HERE WeGo.
Citymapper: Your urban transit companion
- Best for: Public transport in major cities.
- Offline: Save key journeys before travelling. Do not assume that full route planning works underground or without coverage.
- Online: Live multimodal planning, disruptions and real-time departure information.
Tripadvisor: Reviews, now with AI DOWNGRADED
- Best for: Researching destinations, restaurants and attractions.
- AI: Tripadvisor offers an AI trip builder and AI-generated review summaries.
- Offline: We found no current official documentation confirming downloadable offline city maps. Treat it as an online research tool rather than an offline navigator.
MAPS.ME: Global offline navigation, with caveats UPDATED
- Offline: OpenStreetMap-based downloads with turn-by-turn navigation, POI search and contour lines.
- Extras: Hotel booking and eSIM purchases are built into the app.
- Caveat: Pricing and download limits vary by platform and country. Check the current in-app purchase screen before relying on it.
Organic Mapsand CoMaps: Open source, no tracking NEW IN GUIDE
- Offline: Both provide OpenStreetMap-based downloads, offline search and navigation for driving, cycling and walking.
- Privacy: No account is required, and the projects emphasise no ads and no tracking.
- Data quality: Coverage depends on local OpenStreetMap contributors. Trails are often strong; opening hours and smaller POIs can be incomplete.
Cyclists
Komoot: The outdoor enthusiast’s choice UPDATED
- Offline: One region is included free; additional regions can be purchased individually or accessed through Premium.
- AI: Komoot’s ChatGPT integration helps users discover routes from Komoot’s route database. Open the resulting route in Komoot and download it before riding.
- Context: Bending Spoons acquired Komoot in March 2025. Multiple reports estimated that a large share of the original team was laid off after the acquisition; the precise figure was not officially confirmed.
Strava: The athlete’s guide UPDATED
- Offline: Offline maps are available to subscribers.
- AI: Strava’s AI-powered Routes use its Global Heatmap to suggest routes based on where athletes actually run, ride and walk.
- Limitation: Routing and offline-map features are largely tied to a paid subscription.
Ride with GPS: The cyclist’s navigator UPDATED
- Offline: Paid tiers support offline routes, voice cues and cue sheets.
- Newer features: Waypoint alerts, collaborative planning with version history and an MTB trail layer.
- Best for: Bikepacking, long-distance rides and cyclists who need precise navigation.
Motorcyclists
REVER: Route planning made easy
- Best for: Motorcycle touring, community routes and ride tracking.
- Offline: Download routes for use without a signal.
- Extras: Butler Maps roads are available to Pro members.
Scenic: Discover the road less travelled
- Best for: Curated scenic and twisty roads.
- Offline: Offline maps and navigation are available.
- Extras: Rider-specific stops and viewpoints.
Walkers and hikers
AllTrails: For trail hikers UPDATED
- Plus: Includes offline maps and wrong-turn alerts.
- Peak: Adds AI Custom Routes, conditions forecasts, crowd heatmaps and Outdoor Lens plant identification.
- Offline workflow: Create or modify a route online, save it, then download the relevant offline map before hiking.
MapMyWalk: Fitness meets navigation
- Best for: Fitness tracking and walking goals.
- Offline: We could not confirm reliable offline basemaps. Use it alongside a dedicated offline map app.
- Context: MapMyWalk is part of Outside.
OsmAnd: The power user’s offline map NEW IN GUIDE
- Offline: OpenStreetMap-based downloads, turn-by-turn navigation, contours, hillshade and GPX import.
- Paid plans: Maps+ unlocks additional features; Pro adds advanced features such as hourly map updates.
- Pricing: Prices differ by platform, region and billing channel. Check OsmAnd’s current pricing page or in-app store before publication.
AI scorecard
This scorecard is an editorial assessment, not a laboratory benchmark. It rates each app’s main AI feature for usefulness, offline value, reach and trust. “Offline value” measures whether the AI output can lead to something usable without a connection.
AllTrails Peak: Custom Routes 16/20
The clearest path from AI output to an offline route: build it, save it, download it and hike it. The main drawbacks are the Peak subscription and the need to check trail closures and conditions independently.
Komoot: @komoot in ChatGPT 15/20
The ChatGPT integration helps discover real Komoot routes rather than inventing them. To use a result offline, open it in Komoot and download the route.
Google Maps: Gemini navigation and Ask Maps 15/20
Google’s AI features are highly capable but connection-dependent. Ask Maps is initially limited to selected markets, and review-based answers are only as reliable as the reviews behind them.
Strava: AI routes and Athlete Intelligence 15/20
Strava’s routes are informed by its Global Heatmap, and subscribers can use downloaded maps. Most routing and offline features require a subscription.
Waze: Gemini reporting and search 15/20
Waze’s voice reporting is useful, but the app depends heavily on live data. It is not a reliable standalone offline navigator.
Tripadvisor: AI trip builder and review summaries 13/20
Useful for shortlisting places, but review summaries can flatten disagreement and inherit unreliable reviews.
Wanderlog: AI assistant 13/20
Best for refining a plan rather than building one from scratch. Offline access is tied to the paid tier.
Apple Maps: natural-language search 13/20
Natural-language search is convenient, but it is limited to compatible Apple devices and may require a connection. Download maps separately for offline use.
Before you go
- Download maps on Wi-Fi, covering your route plus a margin.
- Download routes and itineraries, not just the base map.
- Test the download in airplane mode before leaving.
- Keep a second fully offline app, such as OsmAnd, Organic Maps, CoMaps or HERE WeGo, as a backup.
- Check closures, access rules, weather and local safety advice separately.
This guide was checked in October 2026. Features, prices and availability can change without notice.
How do you like this article? Read more and subscribe to our monthly newsletter!
#GeoDev
Next article
Can Geography Train AI? When Spatial Relationships Become the Labels
Most GeoAI models still depend on prepared training data. A building detector needs building footprints. A road model needs road centerlines. A land-cover model needs pixel classes. A vision-language model needs captions that explain what appears in an image.
This creates a bottleneck in Earth observation. Satellites, aircraft, and drones keep collecting imagery, but human annotation is slow. Tracing rooftops, drawing roads, marking flood boundaries, or separating crop types often needs domain knowledge. A non-specialist can usually identify a car in a street photo. Labelling a wetland, a seasonal water body, or a mixed crop field from above is harder.
Large datasets show how much effort this requires. Functional Map of the World contains more than one million satellite images from over 200 countries, with bounding-box labels across 63 categories. SpaceNet was built around high-resolution labelled imagery for tasks such as building footprint extraction and road network mapping. DynamicEarthNet combines daily Planet imagery with monthly pixel-level land-cover labels, showing how difficult dense temporal annotation can become.
These datasets are important, but they also show the constraint. High-quality labels are expensive, unevenly distributed, and often tied to specific places or tasks.
This raises the central question of this article: what if some of the information needed to train a GeoAI model already exists in geography itself?
Geography Already Contains Hidden Labels
A satellite image may not have a human label, but it is rarely just a grid of unknown pixels. Its coordinates already place it on the Earth. From that location, we can connect the image to roads, rivers, building footprints, land use, elevation, weather records, and older images of the same place.
These layers are not labels in the usual sense. A digital elevation model does not directly say “urban area.” An OpenStreetMap road line does not describe the full image. A past satellite scene does not replace a human annotation. Still, each one gives the model a clue about the place it is looking at.
Traditional supervision is simple: an image is paired with a human label, and the model learns from that label.
Geographic supervision works differently. The image is paired with spatial context and relationships. A model may learn that a grey roof lies beside a road, that a field sits near an irrigation canal, or that a water boundary follows low-lying terrain. These are weaker than expert labels, but they carry structure that raw pixels alone do not provide.
Early Earth observation and OpenStreetMap experiments showed that map layers can be used together with satellite imagery for semantic labelling. That idea has become more relevant as GeoAI moves toward models that learn from imagery, maps, location, and time together.
Geographical reference imagery is not always the complete ground truth. Maps can be old, incomplete, or misaligned. But they can still act as a source of supervision when labels are limited.
Geographic context layers turning an unlabelled satellite image into a training signal. Source: AI-generated
Five Ways Geography Could Teach a Model
Geography can provide training signals in several ways. These signals are weaker than hand-drawn labels, but they can still help a model learn from the structure around an image.
The first is location. Coordinates are not just numbers. They place an image inside a climate zone, terrain type, settlement pattern, and regional context. SatCLIP learns location embeddings by aligning satellite imagery with geographic coordinates, while GeoCLIP uses location-image alignment for worldwide geolocalization. In both cases, location becomes part of the learning signal.
The second is surrounding context. An unlabelled grey rectangle in an image may be hard to interpret by itself. But if it sits beside a highway, rail yard, port, or irrigation canal, that context gives the model useful clues about what the object might be.
The third is spatial relationship. Geography is not only about what exists, but how things connect. Roads form networks. Buildings sit beside streets. Rivers have upstream and downstream relationships. Sat2Graph uses graph-tensor encoding to extract road graphs from satellite imagery, showing why connectivity matters beyond pixel-level segmentation.
The fourth is time. Satellites revisit the same location repeatedly, creating natural pairs of images from different dates. Seasonal Contrast uses this structure for self-supervised learning, helping models learn what remains stable despite seasonal changes.
The fifth is physical context. Elevation, slope, drainage, coastlines, and hydrology can constrain what is likely or unlikely in an image. A flood prediction on a steep slope, for example, should be treated differently from one in a low-lying floodplain. Physics-guided flood modelling combines remote-sensing imagery with DEM-derived terrain features and hydrodynamic constraints.
Together, these signals suggest a different way to think about supervision. Geography is not only the thing GeoAI tries to map. It can also help teach the model how the world is structured.
This Is Already Starting to Happen
Using geography as a training signal is no longer only a concept. Several recent systems are testing how maps, geographic priors, and spatial graphs can influence how remote-sensing models learn.
These systems are different, but they share one idea: geographic data is beginning to move into the training process itself. It is not only something a model reads after prediction. It can also help form the representation the model learns.
Why This Could Change GeoAI
Geographic supervision could reduce some dependence on manual annotation. Open maps, elevation data, satellite archives, and repeated observations already describe parts of the world at scale. They cannot replace expert labels, but they can give models useful training signals before humans begin drawing polygons. SSL4EO-S12 shows how unlabeled Sentinel-1 and Sentinel-2 archives can support self-supervised pretraining across sensors, seasons, and locations.
It could also make learned representations richer. A model trained only on pixels may learn that a roof has a certain colour or texture. A model trained with geographic context can also learn that the roof sits beside a road, falls inside a residential area, and remains stable across several dates.
Regional transfer may improve, but not automatically. A road in Germany, Kenya, and India may look different, but its network role and relation to nearby buildings may carry useful structure.
This also brings GIS and computer vision closer together. Rasters, vector maps, terrain data, and time-series observations can become part of training, not only post-processing. UN-Habitat’s GeoAI toolkit reflects this broader use of satellite imagery, geospatial data, and planning information together.
GeoAI has mostly learned patterns inside geographic data. It may increasingly learn from the structure of geography itself.
But Geography Alone Is Not Always Ground Truth
Geographic data can be useful, but it can also be incomplete or outdated.
OpenStreetMap is useful because it provides roads, buildings, land use, and other mapped features across many places. But coverage is uneven. One global study estimated the OSM road network to be about 83% complete, while another found strong spatial inequalities in OSM building completeness. . If a training pipeline treats “no mapped building” as “no building exists,” the model may learn false negatives in places where mapping is incomplete.
Spatial distribution of OSM building completeness in 13,189 urban centers. Source: Herfort et al., 2023
Time adds another problem. A road, building, or land-use boundary may remain in a map after the landscape has changed. When an old vector layer is paired with newer imagery, the model receives a conflicting signal.
Coordinates can also become shortcuts. A model may learn that a crop is common in one region instead of learning its visual and spatial characteristics. That can make results look good in nearby test areas but weaker in new regions.
So geographic supervision should be treated as evidence, not truth. The stronger approach combines human labels, imagery, maps, location, time, spatial relationships, and physical context.
GeoAI has usually been trained to learn about geography. The next step may be letting geography itself become part of how the model learns.
Did you like this post? Follow us on our social media channels!
Read more and subscribe to our monthly newsletter!
