Compare data surfaces
| Surface | Best for | What you get | Tradeoff |
|---|---|---|---|
| Standardised formats | Cross-venue analytics, features, and ingestion pipelines | One canonical event envelope across venues | You do not keep venue-native field names |
| Raw payloads | Replayability, parser audits, custom normalization | Unmodified exchange-native messages | You must handle venue-specific schemas yourself |
| Snapshots | Bulk historical ingestion | Daily .jsonl.zst files for standardized datasets | Not the best fit for low-latency recent polling |
| REST endpoints | Recent queries, scripts, product features | Direct API access to trades, events, OHLCV, raw, and metadata | You must page through responses and manage query flow yourself |
Recommended starting points
If you are building a research pipeline
Start with Snapshots and Standardised formats. That gives you the fastest path to a local dataset with a stable cross-venue schema.If you are validating current market activity
Start with Trades API or Events API. Use these when you need recent records rather than bulk historical files.If you are debugging venue behavior
Start with Raw API. Use raw payloads when you need exact source messages or you are maintaining custom parsers.If you are unsure what a market supports
Start with Catalog. It tells you whichexchange and asset pairs are valid and shows historical bounds.
Recommended first path
- Run the Quickstart.
- Confirm access in Connecting.
- Check market availability in Catalog.
- Download one file from Snapshots.
Next steps
- Read Standardised formats if you want one schema across venues.
- Read Raw payloads if you need exact venue-native structure.
- Read API overview if you plan to query recent data directly.