How it works
Platforms name the same idea differently: a follower count is followers_count on one and statistics.subscriberCount on another. Onekey maps every source onto one fixed schema of 147 fields (version 1.0), grouped by what they describe: author, content, engagement, entity and so on. A field a platform does not have comes back as null, never missing. The list here is read from the schema definition the API enforces, and the platform and endpoint lists come from the field mappings of 1,995 endpoints (147 of the fields are returned by at least one). Read more in Unified schema.
Data with no shared meaning across platforms lives under ext.<name>: one flat level, camelCase, documented on the endpoint page that returns it (open an endpoint and look for Field mapping). Today there are 1,236 such names across 1,134 endpoints. They are stable per endpoint but not part of the cross-platform contract, so code that must work on every platform should read the fixed fields only. When a name proves useful on many platforms it is promoted into the fixed schema, additively.