GET
/api/agents/directoryno authSearch a published directory
A directory is a dataset the owner published as a browsable, searchable section — a subsidies list, a catalogue. This is the endpoint behind it, so anything the site can show, an integration can read.
Try it
curl "https://storkie.ai/api/agents/directory?id=YOUR_AGENT_ID&setId=SET_ID&q=solar"A real call to the live API, prefilled with a published agent. Free, and no key.
Query parameters
| Name | Type | Description |
|---|---|---|
idrequired | string | The agent ID. |
setIdrequired | string | Which published directory. |
q | string | Search terms, up to 120 characters. All terms must match. |
slug | string | Fetch a single row by its slug instead of searching. |
Response
| Name | Type | Description |
|---|---|---|
rows | array | Matching rows, with their slugs. |
row | object | When `slug` is given: that one row. |
columns | array | Column names, in order. |
Worth knowing
- Published state is the ONLY gate. A visitor’s own working set — the rows an agent shortlisted for them — is user data and never reachable here.
Rate limit
By IP, shared with other public reads
Errors
400— Missing `id` or `setId`.404— No such agent or directory, or it is not published as one.
Questions
How is this different from the datasets endpoint?
Datasets returns a whole set. A directory is search-first and slug-addressable, which is what makes each row its own indexable page.
Is search fuzzy?
No. Terms are ANDed and matched as substrings, which is predictable — and predictable beats clever when someone is looking for a specific row.