This is a free Research Preview. No SLA, fair-use quotas — one active run at a time — and both the results and the service can change. It predicts one column of one table: two outcomes, several, or a number. Everything it does not do is on the product page, and the preview terms are short.
Three steps
- Register at https://cloud.veox.ai/register. An address is the whole form — no role dropdown, no use-case questions, no call with anyone. A priority access code admits you straight away; without one you join the waitlist, which is currently 1–3 weeks, and we write to you when it is your turn. It is free either way.
- Upload a table, or start from an example. A CSV or TSV of your own, or one of the example datasets if you want to see the shape of it first. Then say which column you want predicted. What we accept covers the formats and the limits.
- Watch it search, then take the result. The leaderboard moves while it works. When it finishes you can export the model, the report and the evidence receipt.
What actually happens
jAIn profiles the table, infers which of the three tasks it is — binary classification, multiclass classification or regression — then searches: it engineers features, tries models against each other, and keeps the one that wins on data it was never fitted on. You get the winner rather than an endpoint to call.
Where to go next
- Want it on your own machine instead of the browser? Running locally installs the same workspace in Docker.
- Want to script it? Drive it from Python does the whole loop over HTTP, and shows how to run the winning model yourself.
Your file is deleted automatically after 30 days. What the run made from it — your report, your trained model and its evidence receipt — is kept. See what we keep.