Tools

The server exposes four generic tools. They are not per-resource — instead, each one takes a resource argument, so the same four tools cover all 77 resources. Two are for discovery, two are for reading data.

Tool	Purpose
`api_v3_resources`	List every resource available, with a short description.
`api_v3_resource_schema`	Describe one resource: attributes, filters, operators, relationships.
`api_v3_list`	Fetch a paginated list of records from a resource.
`api_v3_get`	Fetch a single record by id.

A typical agent calls them in that order: discover what exists, learn how to query it, then query.

`api_v3_resources`

Lists every read-only resource exposed via MCP. It's a lightweight index — call it once at the start of a session to learn what exists.

Arguments: none.

Returns an array of resources, each with a name, a one-line description, and a filterable flag (whether the resource supports attribute filters):

{
  "resources": [
    {
      "name": "bookings",
      "description": "Reservations and unavailable periods. By default returns visible (non-canceled) bookings.",
      "filterable": true
    },
    {
      "name": "bookings_fees",
      "description": "Per-booking fees attached to reservations (cleaning, tax, etc.).",
      "filterable": true
    },
    {
      "name": "rentals",
      "description": "Listings/properties for the account.",
      "filterable": true
    }
  ]
}

To learn a specific resource's attributes and filters, follow up with api_v3_resource_schema.

`api_v3_resource_schema`

Returns the detailed schema of a single resource so the agent knows exactly what it can request and filter on — without trial and error.

Arguments:

Argument	Type	Required	Description
`resource`	string	yes	Resource name (from `api_v3_resources`), e.g. `bookings`.

Returns an object describing:

attributes — every field in the response (this doubles as the response shape). Each has a type, a value format hint, a filterable flag, and the comparison operators it accepts.
relationships — associated resources, and (for belongs_to) the filter keys/type/operators to filter by them; polymorphic relationships list a companion requires key.
resource_filters — resource-specific filters accepted as top-level arguments (e.g. bookings' status, from, until).
standard_filters — ids, updated_since, limit, offset (available on every resource).
sparse_fieldsets — always true.
filter_examples — ready-to-use filter payloads for this resource.

Example (excerpt) for bookings:

{
  "name": "bookings",
  "description": "Reservations and unavailable periods. By default returns visible (non-canceled) bookings.",
  "attributes": [
    { "name": "id",        "type": "integer",  "format": "integer",  "filterable": true,  "operators": ["eq","not_eq","gt","gteq","lt","lteq"] },
    { "name": "start_at",  "type": "datetime", "format": "ISO 8601", "filterable": true,  "operators": ["eq","not_eq","gt","gteq","lt","lteq"] },
    { "name": "end_at",    "type": "datetime", "format": "ISO 8601", "filterable": true,  "operators": ["eq","not_eq","gt","gteq","lt","lteq"] },
    { "name": "status",    "type": "computed", "format": null,       "filterable": false, "operators": [] },
    { "name": "reference", "type": "string",   "format": "string",   "filterable": true,  "operators": ["eq","in","not_eq","matches","does_not_match"] }
  ],
  "relationships": [
    {
      "name": "rental",
      "kind": "has_one",
      "polymorphic": false,
      "resource": "rentals",
      "filter": { "keys": ["rental_id", "rental"], "type": "integer", "operators": ["eq","not_eq","gt","gteq","lt","lteq"] }
    }
  ],
  "resource_filters": [
    { "name": "status", "type": "string", "description": "Comma-separated: booked, unavailable, tentative" },
    { "name": "from",   "type": "string", "description": "Only bookings starting on/after this ISO 8601 date (default: now)" }
  ],
  "standard_filters": ["ids", "updated_since", "limit", "offset"],
  "sparse_fieldsets": true,
  "filter_examples": [
    { "reference": "..." },
    { "start_at": { "op": "gt", "value": "2026-01-01T00:00:00Z" } },
    { "start_at": [ { "op": "gteq", "value": "2026-01-01T00:00:00Z" }, { "op": "lt", "value": "2026-12-31T00:00:00Z" } ] },
    { "rental_id": 1 }
  ]
}

computed attributes

Attributes with "type": "computed" are derived values that aren't backed by a database column. They are returned in the response but cannot be filtered on (filterable: false, empty operators).

See Querying for how to turn this schema into actual filter and fields arguments.

`api_v3_list`

Fetches a paginated list of records from a resource. Returns payloads in the same shape as the API v3 index endpoint.

Arguments:

Argument	Type	Required	Description
`resource`	string	yes	Resource name (from `api_v3_resources`).
`account_id`	integer	conditional	Account to operate on. Required for multi-account tokens, ignored for single-account ones. See account selection.
`ids`	string	no	Comma-separated list of IDs to fetch, e.g. `"12,34,56"`.
`updated_since`	string	no	ISO 8601 timestamp; only records updated after this time.
`limit`	integer	no	Page size (default 25, max 100).
`offset`	integer	no	Pagination offset (default 0).
`filter`	object	no	Attribute/relationship filters. See Querying → Filtering.
`fields`	array	no	Sparse fieldset: return only these attributes. See Querying → Sparse fieldsets.

Resource-specific filters (such as bookings' status, from, until) are passed as top-level arguments alongside the standard ones — api_v3_resource_schema lists them under resource_filters.

Example call (bookings for one rental, two fields, first page):

{
  "name": "api_v3_list",
  "arguments": {
    "resource": "bookings",
    "filter": { "rental_id": 42 },
    "fields": ["start_at", "end_at", "final_price"],
    "limit": 10,
    "offset": 0
  }
}

Response shape — a key named after the resource, plus a meta object:

{
  "bookings": [
    { "id": 123, "start_at": "2026-07-01T00:00:00", "end_at": "2026-07-08T00:00:00", "final_price": "1450.0" }
  ],
  "meta": { "total_count": 1, "limit": 10, "offset": 0 }
}

`api_v3_get`

Fetches a single record by id. Returns the payload in the same shape as the API v3 show endpoint.

Arguments:

Argument	Type	Required	Description
`resource`	string	yes	Resource name (from `api_v3_resources`).
`id`	integer	yes	The record ID.
`account_id`	integer	conditional	Account to operate on. Required for multi-account tokens, ignored for single-account ones. See account selection.
`fields`	array	no	Sparse fieldset: return only these attributes (`id` always included).

Example call:

{
  "name": "api_v3_get",
  "arguments": {
    "resource": "bookings",
    "id": 123,
    "fields": ["start_at", "end_at", "final_price"]
  }
}

Fetching many records by id

To fetch several specific records in one round-trip, prefer api_v3_list with the ids filter ("ids": "12,34,56") over multiple api_v3_get calls.

How results are returned over MCP

Per the MCP spec, a tools/call result wraps the JSON payload as text content. The JSON shown above is what you'll find inside content[0].text:

{
  "content": [
    { "type": "text", "text": "{\"bookings\":[{\"id\":123, ...}],\"meta\":{\"total_count\":1,\"limit\":10,\"offset\":0}}" }
  ]
}

MCP clients parse and present this for you; you rarely need to handle the envelope directly. See Protocol & transport for the raw JSON-RPC details.

api_v3_resources​

api_v3_resource_schema​

api_v3_list​

api_v3_get​

How results are returned over MCP​

`api_v3_resources`

`api_v3_resource_schema`

`api_v3_list`

`api_v3_get`

How results are returned over MCP