Examples

End-to-end walkthroughs: a real question, the commands that answer it, and the output. Each one runs against the live database, so the numbers in the page are numbers the CLI actually returned.

They're written for use with Claude Code or any agent that can run a shell. The pattern is always the same, and it's the pattern worth copying:

fetch to disk → look at what came back → build.

The middle step is the one people skip. Every example here shows what it caught.

Available now

Competitive landscape: TL1A

The full walkthrough. Three commands and a script produce a landscape spreadsheet, a bullseye and the supporting statistics for a crowded target: 56 assets, 66 trials, 59 deals, three Phase 3 programmes all in ulcerative colitis and all acquired rather than built.

Covers paging past the per-entity --limit cap, and the two-minute check worth running on any cohort before charting it.

Suggest one

These are meant to cover the questions people actually bring to the data. If the one you need isn't here, say so, it's a good signal for what to write next.