I'm Nazim, founder of Koinju. We've been ingesting per-exchange trade data for many years (now 286B trades, 25 exchanges) for an institutional product. We just opened our DB via direct SQL access. We offer a Free tier: 50 queries/month to test the product.
I'd love any feedback on SQL vs REST framing. Most of what people write in pandas/polars scripts to aggregate trades for something that could be very simple to do in direct SQL. Cross-exchange index calculation went from 2,500 lines of Python to 7 lines of SQL. Multi-exchange OHLCV: 130 lines to 3. And we added multiples exemples in our doc : https://docs.koinju.io/compute-engine/introduction
Happy to answer anything, schema, query plan, ingestion architecture, why sql not rest, what we don't have, what we got wrong.
Thanks!