Delete the rows matching a filter

curl --request DELETE \
  --url https://api.paigeme.dev/v1/tables/{table}/rows \
  --header 'Authorization: Bearer <token>'

const options = {method: 'DELETE', headers: {Authorization: 'Bearer <token>'}};

fetch('https://api.paigeme.dev/v1/tables/{table}/rows', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "success": true,
  "data": {
    "table": "<string>",
    "schema": "<string>",
    "deleted": 123
  }
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

Delete the rows matching a filter

Deletes every row matching filters and returns how many went. The rows themselves are not echoed back.

filters is required. There is no “delete everything” form of this endpoint — omitting the parameter is 400, never a whole-table wipe. To clear a table, use the dashboard Database tab.

Predicates travel in the query string as a JSON array (a DELETE body is widely stripped in transit, and a dropped body here would be catastrophic): filters=[{"column":"id","op":"eq","value":42}]. AND-joined, 1-20. Ops: eq, neq, gt, gte, lt, lte, is.

like and ilike are read-only and are rejected here (400 filter_op_not_allowed). A pattern match has no bounded row set, so Paige will not delete through one — read the rows you mean from GET /v1/tables/{table}/rows first, then delete them by primary key.

Every filter column is checked against the table’s real columns (400 unknown_column), and a column colliding with a reserved database query parameter is refused (400 reserved_filter_column) because its predicate would be silently discarded.

Defaults to the public (production) schema.

A Paige platform table is 403 platform_table_forbidden; an unknown table is 404 table_not_found.

Only base tables are addressable. A view (or any other non-table relation) answers the same 404 table_not_found an unknown name does — including a view defined over a Paige platform table, which would otherwise be a writable alias for it.

Required scope: tables:write

DELETE

tables

{table}

rows

Delete the rows matching a filter

curl --request DELETE \
  --url https://api.paigeme.dev/v1/tables/{table}/rows \
  --header 'Authorization: Bearer <token>'

const options = {method: 'DELETE', headers: {Authorization: 'Bearer <token>'}};

fetch('https://api.paigeme.dev/v1/tables/{table}/rows', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "success": true,
  "data": {
    "table": "<string>",
    "schema": "<string>",
    "deleted": 123
  }
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

{
  "success": false,
  "error": {
    "code": "invalid_request",
    "message": "Invalid request body",
    "details": {}
  },
  "request_id": "<string>"
}

Authorizations

Authorization

string

header

required

Project API key issued in Settings → API keys. Send it as Authorization: Bearer pk_live_….

Headers

X-Paige-Project

string

Target project id, for an MCP OAuth bearer (mcp_at_…) attached to more than one project — get ids from GET /v1/projects. Matched case-insensitively.

Omit it and a READ falls back to the connection's default project; a mutation (any non-GET) on a connection with 2+ projects is rejected with 400 project_required — a write is never defaulted to a guessed project. A connection with exactly one project never needs the header.

Scopes are checked against the SELECTED project only, never a union across the connection.

For a pk_ API key the header selects nothing — one key is one project's context — but it IS validated: omit it and the key's own project is used, send it and it must name that project, otherwise the call is rejected with 403 project_not_attached (a blank value is 400 invalid_project_header, as above).

Path Parameters

table

string

required

Table name. Lowercase letters, digits and underscores only. Paige platform tables are rejected.

Pattern: ^[a-z_][a-z0-9_]{0,62}$

Query Parameters

schema

enum<string>

Which database schema to target: "public" (production, the default) or "preview" (dev sandbox).

Available options:

public,

preview

filters

string

required

REQUIRED. A JSON array of predicates, e.g. [{"column":"id","op":"eq","value":42}]. AND-joined, 1-20. There is no "delete everything" form.

Required string length: 1 - 4000

Response

Success.

success

enum<boolean>

required

Available options:

true

data

object

required

Hide child attributes

data.table

string

required

data.schema

string

required

data.deleted

number

required

How many rows matched the filters and were deleted.

Getting started

Endpoints

Delete the rows matching a filter

Authorizations

Headers

Path Parameters

Query Parameters

Response