How to Convert Your Google Timeline to GPX (Records.json or Semantic Location History)

Google spent 2024 and 2025 dismantling the Timeline its users loved: on-phone storage replaced the cloud copy, retention settings started silently trimming the history, and the desktop Timeline view went away. What remains accessible is a Takeout export — and a growing number of people grabbing theirs before the retention clock finishes.

The catch: what Takeout gives you is JSON, in one of two shapes, and almost nothing displays it. This guide walks the whole path: exporting the right archive, understanding which file you got, and turning it into a GPX track that opens in any mapping app — with the file never leaving your machine.

Step 1: Export from Google Takeout

1. Go to takeout.google.com and sign in.
2. Click Deselect all — you want one product, not your digital life.
3. Tick Location History (Timeline).
4. Under "Multiple formats", note the choice offered: JSON is what you want (KML exports a coarse track, not the history).
5. Create the export. Depending on the account, Google mails a link or offers a direct download; the archive is a zip.

Unzip it. Inside you'll find one of three things:

• A single Records.json — the current all-in-one export holding your entire remaining history.
• A Semantic Location History folder with one JSON per month (2023_OCTOBER.json and friends) — the older, still-available format.
• A Location History.json — the classic export from a few years ago, a single file with a plain location stream. If you downloaded your history back then and still have the file, it converts too.

Any of them works. The converter on this site detects which one you have from the file's structure, so there's no configuration step.

Step 2: Convert to GPX

Open the Google Timeline to GPX converter and drop the JSON onto it. A few seconds later you get a GPX file — the universal GPS format that opens in Google Earth, QGIS, every GPX viewer, and anything else that draws tracks.

What the conversion does under the hood:

• Movement becomes track. Each activity segment — walking, driving, cycling — becomes a track segment with timestamps. A day with a commute, a lunch walk and a drive home arrives as three ordered segments.
• Visits are skipped. A place visit is a location with a duration, not a movement; converting it would just stack duplicate points. They're left out, deliberately.
• Coordinates are cleaned. Google's data uses E7 integers (degrees × 10,000,000) and includes (0,0) placeholders for missing fixes; both are handled, along with consecutive duplicates.
• Everything happens locally. This is the file with your home address in it. The tool parses it in your browser — no upload, verifiable in the network inspector.

One practical warning about size: a long history can make Records.json enormous — hundreds of megabytes of JSON. The converter runs in a background worker and handles big files, but if your machine struggles, choose the Semantic Location History export instead and convert month by month. Same data, friendlier chunks.

Step 3: Use it

A recovered Timeline as GPX goes wherever tracks go:

• Google Earth — drop the GPX in and see years of movement over real terrain. (Or convert GPX to KML first for Earth-native styling.)
• QGIS or a web map — convert to GeoJSON and your history becomes ordinary data: heatmaps, filters, measurements.
• A personal heatmap — drop the GPX (and any fitness files) into the heatmap generator and watch your actual life-pattern render.
• An archive — GPX is plain text and will open in twenty years regardless of what Google does next. For the most sensitive dataset you own, a durable local copy is the point.

Accuracy expectations

Be kind to the data: Google Timeline positions are inferred from phone sensors, Wi-Fi and cell towers. They're good to city-block scale, occasionally better downtown, worse in the country. The converter renders exactly what Google recorded — no smoothing, no invented positions — so a stray jump in the track is Google's record, not a conversion bug. Treat the result as a memoir, not a survey.

The move that matters is Step 0: export before retention settings eat the history. Everything after that is one drag and one download.