> ## Documentation Index
> Fetch the complete documentation index at: https://docs.plyne.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Free parameter helpers

> Discover exact provider filter values through the API and MCP without spending Plyne credits.

Use the same workspace API key as acquisition tools. Helpers are read-only, cost
0 Plyne credits, require no idempotency key and never execute an acquisition tool.
Authentication and workspace rate limits still apply.

## Discover catalogs

API: `GET /v1/catalogs`. MCP tool: `list_catalogs` with `{}`.
The response lists catalog IDs, namespaces, descriptions, availability, applicable tool/field paths and credits.
A source that is not configured is marked unavailable; it does not return invented
options. The first release supports countries, LinkedIn V2 industries, TheirStack
industry IDs from the checked-in V2 reference, FullEnrich industry labels and
TheirStack keywords. Industry reference membership is not proof that every
provider endpoint supports a value or that a company matches it.

## Search or verify values

API: `GET /v1/catalogs/{catalog}/options?q=...&limit=20&page=0`.
MCP tool: `search_catalog`:

```json theme={null}
{
  "catalog": "theirstack.keywords",
  "query": "SAP",
  "filters": { "keyword_type": "technology" },
  "limit": 20
}
```

Copy exact returned values into the acquisition field. Preserve case and namespace.
A suggestion is not an automatic identity match. Use next\_page when present.
Query length is at most 200 characters, limit is 1–50 and page is 0–100.

To verify previously selected IDs, use MCP values (an array of up to 50 exact
strings), or repeat the API value query parameter. Set limit large enough for
all selected values. Do not combine values with query, filters or a nonzero page.
Unknown IDs are returned in unresolved\_values without fuzzy substitution.

Keyword filters are keyword\_type, category\_slug and parent\_category\_slug. In the
API they are query parameters; in MCP they are nested under filters. TheirStack
keyword search needs a configured provider. Lookup pages reuse the existing
catalog cache. Customer acquisition credits are not reserved.

## Inspect a field

API: `GET /v1/tools/job_search/field-options?provider=theirstack&field=company_technology_slug_or`.
MCP tool: `get_field_options`:

```json theme={null}
{
  "tool": "job_search",
  "provider": "theirstack",
  "field": "company_technology_slug_or"
}
```

The response contains the field schema, fixed options where an enum exists, and
a catalog reference when supported. With provider, field may be relative to that
provider's namespace. Otherwise use the complete dot-separated input field path.
A field without options is not necessarily arbitrary free text: read its
constraints and description. Entire object schemas require selecting a leaf field.

Executable input schemas also carry x-plyne-catalog references, including the
catalog ID and search\_catalog helper name. These appear in API descriptors,
MCP resources and the portal reference. Helpers do not broaden fallback eligibility:
a provider-specific filter still selects that provider.

City/region lookup, entity typeahead and unreviewed provider-specific vocabularies
are not included. Existing paid count or identity tools remain separate.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.