search

[Search and Aggregates]

Find ranked current documents. Auto uses lexical ranking; hybrid needs --semantic and a ready index. Call describe after a hybrid failure. Use list_documents for complete sets.

Parameters

Name	Type	Required	Description
query	string	Yes	Search text: one to 32 distinct terms.
spaceId	string	Yes	Workspace ID from list_spaces.
collectionIds	json	No	Search only these collections of the workspace.
cursor	string	No	Opaque token from a prior page. Unknown tokens require enumeration to restart.
filter	json	No	Typed conditions, all of which must match: {"all": [...]}, up to 16. Each has a path (a document root such as /title or /tags, or a /data pointer), an op, and a type. exists needs only path and op. eq, contains, gt, gte, lt, and lte take type (string, number, boolean, null, date, or instant) and value; in takes type and values (up to 50).
includeArchived	boolean	No	Include archived records. Default: `false`
limit	number	No	Maximum items per page, from 1 to 50. Default: `20`
mode	string	No	Auto selects lexical. Explicit hybrid needs --semantic and a ready index. Call describe after a retryable failure. Default: `"auto"`

Examples

Replace the sample IDs with values from your store. Use a fresh UUID for each requestId.

Find documents about the deployment window.

Arguments 
{
  "spaceId": "11111111-1111-4111-8111-111111111111",
  "query": "deploy window",
  "limit": 5
}

Full MCP request

Full MCP request 
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "search",
    "arguments": {
      "spaceId": "11111111-1111-4111-8111-111111111111",
      "query": "deploy window",
      "limit": 5
    }
  }
}