Querying

All data in Micro is queried through a single endpoint:

POST /v2/prism/{teamId}/{objectType}/query

The request body contains a query object that describes what to fetch.

Select

select is required. Pass an array of property slugs to return.

{
  "query": {
    "select": ["full_name", "email", "title"]
  }
}

Use dot notation to traverse relationships:

{
  "query": {
    "select": ["full_name", "company.name", "company.primary_domain"]
  }
}

Filter

filter is an array of conditions. Each condition is { slug: { operator: value } }.

{
  "query": {
    "select": ["full_name", "email"],
    "filter": [
      { "labels": { "in": ["investor"] } },
      { "last_interaction_date": { ">=": "2024-01-01" } }
    ]
  }
}

Operators

Operator	Types	Description
=	all	Exact match
!=	all	Not equal
< > <= >=	number, date	Comparison
in	string, select	Value is in the array
not_in	string, select	Value is not in the array
begins_with	string	Starts with value
ends_with	string	Ends with value
contains	string	Contains value
not_contains	string	Does not contain value
between	number, date	Within a two-element [min, max] array, inclusive on both ends
exists	any	Field has a value
not_exists	any	Field has no value
is_null	any	Alias for not_exists — takes any truthy value
is_not_null	any	Alias for exists — takes any truthy value

Each filter object takes exactly one property with exactly one operator.

Combining filters

By default filters are combined with AND. Use combinator to switch to OR:

{
  "query": {
    "select": ["full_name"],
    "filter": [
      { "labels": { "in": ["investor"] } },
      { "labels": { "in": ["founder"] } }
    ],
    "combinator": "OR"
  }
}

Sort

sort is an array of { slug: direction } objects. Earlier items in the array take priority.

{
  "query": {
    "select": ["full_name", "last_interaction_date"],
    "sort": [
      { "last_interaction_date": "desc" },
      { "full_name": "asc" }
    ]
  }
}

Pagination

Use limit to set page size. It must be between 1 and 50 — requests above the cap are rejected rather than clamped.

{
  "query": {
    "select": ["full_name"],
    "limit": 50
  }
}

The response wraps rows in a data array alongside pagination metadata:

{
  "data": [
    {
      "id": "c1a2b3c4-...",
      "properties": {
        "full_name": "Sarah Chen",
        "email": "sarah@example.com"
      }
    }
  ],
  "has_more": true,
  "next_cursor": "eyJrIjoi..."
}

To fetch the next page, pass next_cursor back unchanged as cursor. When cursor is set, page and limit are derived from the cursor and any explicit values are ignored. has_more is false on the last page, at which point next_cursor is null.

{
  "query": {
    "select": ["full_name"],
    "cursor": "eyJrIjoi..."
  }
}

Set include_total: true (a sibling of query, not inside it) to add an unpaginated total to the response. It costs an extra pass over the result set, so prefer GET /v2/prism/{teamId}/{objectType}/count for unfiltered totals.

:::caution page still works but is deprecated — page-number pagination drifts under concurrent writes. Use cursor for anything other than a one-shot export. :::

Scoping to a list

Pass list_id — a UUID — to scope results to a specific list (e.g. your fundraising pipeline vs. your investor network):

{
  "query": {
    "select": ["full_name", "stage"],
    "list_id": "3f9a1c42-5b7e-4d18-9a2f-8c1e6b0d7f35"
  }
}

query rejects unknown properties, so a misspelled or renamed key returns 400 rather than being silently ignored.