Use the documented dataset schemas
NexusTrade lake queries use canonical lake.* schemas backed by manifest-resolved data. The catalog provides columns and semantic guidance so scripts and agents reference the same logical contracts. Read-only access is separate from portfolio mutation and live-order approval.
Save the query and its research inputs
Retain the dataset schema, query, result manifest, source dates, exact strategy configuration and operation IDs. Compare historical tests on the same data and execution assumptions. Missing or stale source values remain explicit instead of being rewritten into confident metrics.
Choose an access method
Use the Python or TypeScript SDK for lake queries and historical backtests, or connect an MCP client for research tools authorized by your account. Enterprise SSO, custom SLAs, institutional licensing and dedicated environments are not established product capabilities.
Check your account scopes and approve account actions in the signed-in app.