Hinge export guide · checked 30 October 2026
Hinge CSV exports: what changed and what you can measure
A Hinge download can be a folder-based CSV export instead of the older JSON layout. If you cannot find matches.json, the download may still be valid. The important question is what the tables actually contain.
Recognize the table-based format
In the layout supported by Pavonia’s CSV importer, tables can appear in folders such as match_likes and chat_message, each containing a CSV file. This is different from an older archive organized around user.json and matches.json.
Open the original ZIP in Pavonia. The importer detects the supported layout. Renaming a CSV file to JSON will not convert its contents or create the relationships missing from the export.
What the supported tables can tell you
| Exported information | What Pavonia can use it for |
|---|---|
| Likes and roses sent | Activity over time and whether a sent like included a comment, when those fields are present. |
| Passes | Counts of recorded passes in the supported skip table. |
| Messages with conversation identifiers | Conversation activity, using the first recorded message to place a conversation on the timeline. |
| App request logs and time-zone history | Additional activity and local-day information when the relevant tables are available. |
These are capabilities of the supported importer, not a promise that every Hinge archive includes every field. The import summary and the report’s coverage notes are the place to check your own file.
Why conversations per like is not your exact match rate
In the supported CSV layout, a sent like does not identify the match it led to. Conversations provide evidence of matches you wrote in, but do not establish a complete count of all matches or their creation times.
For that reason, Pavonia labels the relevant measure conversations per like when complete match coverage is absent. It counts conversations you wrote in, on the day of the first recorded message, relative to likes sent. A match with no recorded message may be missing from that numerator.
Do not read that ratio as the probability that one particular like will produce a match. Conversations and likes in the same period are not necessarily the same set of interactions. The CSV layout also cannot establish which individual commented likes succeeded when their outcomes are absent.
Compare like with like
Check the export format, date coverage and missing fields before comparing reports. A lower conversation count can reflect a different export layout or messaging behavior rather than a change in the number of people who matched with you.
- Use comparable date ranges.
- Keep matches and conversations labeled separately.
- Treat missing outcomes as unknown, rather than failed likes.
- Read warnings about missing timestamps, records and unsupported calculations.
A partial archive should lead to a qualified result, not a more confident-looking number. Pavonia’s report keeps coverage limitations beside the affected measures.
Get a fresh export or inspect your existing one
Follow the Hinge data-download guide if you need a new archive. Hinge’s official help page explains its request process and notes that exports do not provide every category of data.
Your imported files are processed locally in the browser. Website traffic and fixed usage events are separate from the dating data in your report; see the traffic privacy notice for those analytics.