Skip to content

Location

Google Maps Timeline and Location History

Google Location History and Maps Timeline: the legacy cloud Records.json, the 2024 to 2025 move to on-device storage, Timeline.json exports and what remains.

Location
On device (2024+): inside the Google Maps and Google Play services sandboxes under /data/data/; exports: Timeline.json; legacy cloud: Takeout Location History (Records.json, Semantic Location History)
Proves
Where the device (and the signed-in Google account) was over time: raw position fixes with accuracy, inferred place visits with start and end times, and travel segments with activity type
Timestamps
Takeout Records.json: ISO 8601 UTC strings (older exports: timestampMs in Unix milliseconds); Timeline.json: ISO 8601 with UTC offset; semantic history: start and end timestamps per visit or segment
Access
On-device Timeline: full file system (root, exploit or agent), credential-encrypted sandboxes, AFU only. User export from the phone's Timeline settings. Legacy cloud history: Google Takeout or legal process to Google, only for data not yet migrated or deleted.
Retention
User-controlled auto-delete (3, 18 or 36 months, or never); on-device data is lost with the phone unless the encrypted Timeline backup was enabled; legacy cloud history was deleted by Google after the migration deadlines
Collection
Full file system extraction, Timeline export from the phone (user), Google Takeout (legacy), Legal process to Google (legacy)

What it is

Google Location History was, for more than a decade, the richest location source tied to an Android phone: the device reported positions to Google, which stored them in the account and rendered them as the Maps Timeline. Because the data lived in the cloud, investigators obtained it through Google Takeout or legal process rather than from the handset.

That changed in 2024 and 2025. Google renamed the setting Timeline, moved storage to the device, and asked users to migrate their history before deadlines that ran into 2025; unmigrated cloud history was then deleted. Users can opt into an end-to-end encrypted cloud backup of their on-device Timeline. For current cases the handset is now the primary source, and the phone must be extracted to reach it.

Where it lives

EraSourceFormat
Legacy (until 2024 to 2025)Google account, exported through Takeout: Location History/Records.jsonJSON array of raw fixes: latitudeE7, longitudeE7, accuracy, timestamp, source, deviceTag, activity guesses
LegacyTakeout: Location History/Semantic Location History/<year>/<year>_<MONTH>.jsontimelineObjects: placeVisit and activitySegment with durations
On-device (2024+)Google Maps (com.google.android.apps.maps) and Google Play services (com.google.android.gms) sandboxesApp-private databases and protobuf blobs; file names and layout have changed between app versions
On-device exportSettings > Location > Location services > Timeline > Export Timeline dataTimeline.json with semanticSegments, rawSignals and userLocationProfile

Public research on the on-device store is still maturing. Identify the files on your own extraction by searching both sandboxes for recently modified databases and for known place names, and validate against a test device on the same app versions.

What it proves

  • The device's position at specific times, with an accuracy radius per point, from GPS, Wi-Fi or cell sources.
  • Inferred visits to places (with Google's place ID and sometimes name and address), with start and end times and a confidence value.
  • Travel between places with an inferred activity (walking, cycling, in a vehicle, on a train) and distance.
  • In legacy data, which device reported each point (deviceTag), useful when an account had several phones.

It does not prove who carried the phone. Inferred visits and activities are model outputs, not measurements: cite the raw points and their accuracy when location matters.

Key fields

FieldWhereMeaning
latitudeE7, longitudeE7Records.jsonCoordinates multiplied by 10^7 (divide to get degrees)
accuracyRecords.jsonRadius in metres
sourceRecords.jsonGPS, WIFI, CELL and similar
placeVisit.duration.startTimestamp / endTimestampSemantic historyVisit window
activitySegment.activityTypeSemantic historyInferred mode of travel
semanticSegments[].visit.topCandidateTimeline.jsonVisited place with placeId, semanticType, probability
semanticSegments[].timelinePathTimeline.jsonPoints along a route (point as "lat°, lng°", time)
rawSignals[].positionTimeline.jsonRaw fixes with LatLng, accuracyMeters, timestamp

Timestamps

Current exports use ISO 8601 strings with an explicit offset (2026-09-14T17:43:12.000+02:00): convert to UTC before merging. Older Records.json exports used timestampMs, Unix milliseconds as a string.

# Legacy Records.json: points on one day, UTC, degrees
jq -r '.locations[] | select(.timestamp | startswith("2023-06-14"))
  | [.timestamp, (.latitudeE7/1e7), (.longitudeE7/1e7), .accuracy, .source] | @csv' Records.json
# Timeline.json: place visits with start and end
jq -r '.semanticSegments[] | select(.visit) | [.startTime, .endTime, .visit.topCandidate.placeId] | @csv' Timeline.json

Retention

The user chooses auto-delete after 3, 18 or 36 months, or keeps everything, and can delete days or ranges manually. On-device Timeline disappears with the device or a factory reset unless the encrypted backup was on and the user restores it. Legacy cloud history that was not migrated has been deleted by Google; old Takeout archives held by the user or by a previous case may be the only remaining copy.

Collection

  • Full file system extraction (AFU) of both the Maps and Play services sandboxes.
  • Export from the phone with the user's cooperation, documented on video if possible.
  • Takeout or legal process for historic cloud data, noting that Google may no longer hold it.

Parsing

  • ALEAPP parses several on-device Google location artifacts; check its module list against your app versions.
  • RLEAPP parses Google Takeout and legal-process returns, including Location History.
  • jq handles both JSON formats for quick filtering and CSV export.
  • Commercial suites parse Takeout JSON and some on-device stores.

Investigator tips

  • Corroborate a visit with independent sources: Wi-Fi networks joined at the place, photo EXIF GPS, and Bluetooth pairings such as a car head unit.
  • Check that the Google account holding the history is the one on the device in the accounts databases.
  • Large accuracy radii (hundreds of metres) usually mean cell-based fixes; do not place someone in a specific building from them.
  • The location forensics guide explains how to weigh these sources together.

See also