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

# monday.com

> Read boards, items, updates, documents, and account data from monday.com

The monday.com source reads work records and account data through monday.com's
GraphQL API. It uses an
[HTTP manifest](/pages/connectors/building-a-connector/http-manifests) and is
**alpha**. Queries have been checked against the published API schema. The
default resources and item and document pagination have been tested with a
small live account; non-empty audit incremental reads remain unverified.

## Configuration

| Field       | Scope      | Default | Description                                      |
| ----------- | ---------- | ------- | ------------------------------------------------ |
| `api_token` | Connection |         | Required. Secret. Personal monday.com API token. |

Find your token under **Developers → API token → Show** in your profile menu.
The token reads whatever that user can access. See
[monday.com's authentication guide](https://developer.monday.com/api-reference/docs/authentication).
Connection testing makes one account query; it does not verify access to every
optional resource.

## Resources

| Group            | Resources                                                                                                                                       |
| ---------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- |
| Account          | `account`, `users`, `teams`, `account_roles`, `departments`                                                                                     |
| Workspaces       | `workspaces`, `folders`, `workspace_users`, `workspace_owners`, `workspace_teams`, `workspace_team_owners`                                      |
| Boards           | `boards`, `archived_boards`, `deleted_boards`, `columns`, `groups`, `views`, `board_team_owners`, `board_team_subscribers`, `board_validations` |
| Items            | `items`, `subitems`, `item_assets`, `subitem_assets`, `item_description_blocks`, `subitem_description_blocks`                                   |
| Updates          | `updates`, `board_updates`, `update_viewers`                                                                                                    |
| Documents        | `docs`, `doc_blocks`                                                                                                                            |
| Reference data   | `tags`, `managed_columns`, `object_schemas`, `directory_resources`                                                                              |
| Product features | `notetaker_meetings`, `custom_activities`, `timeline_items`, `subitem_timeline_items`, `sequences`                                              |
| User settings    | `favorites`, `notifications`, `notification_settings`, `board_notification_settings`                                                            |
| Administration   | `webhooks`, `audit_logs`, `audit_event_catalogue`, `api_usage`                                                                                  |

Selected by default: `account`, `users`, `teams`, `workspaces`, `folders`,
`boards`, `columns`, `groups`, `views`, `items`, `subitems`, `updates`, `docs`,
`doc_blocks`, and `tags`. The rest are optional and may require additional
permissions, a paid plan, or access to the corresponding product. Audit logs
require Enterprise admin access and the `manage_account_security` permission.

Selecting a child resource fetches its parents automatically. Parent rows are
only written when also selected. IDs are strings; child records use parent IDs
in their keys where needed. Nested structures remain JSON.

`items` reads active top-level items on classic and multi-level boards.
`subitems` reads their descendants with parent and board IDs. Item rows include
column definitions and values, but evaluated formula values are not requested.
Archived and deleted boards contain metadata only; their items are not read.

Updates include replies, reactions, and file metadata. Optional `board_updates`
adds board IDs to board-level posts, which can also appear in `updates`.
Document and item-description blocks are read separately. Assets contain
metadata, not file contents.

Not included: board and user activity histories, dashboards and workflows,
knowledge-base articles, WorkForms definitions, app-developer APIs, document
version history, and reply viewers. Webhook configuration is read without
creating subscriptions or receiving events.

## Modes

All resources support full reads. Only `audit_logs` supports incremental reads.

| Resource     | Cursor field | Request parameter        | Checkpoint key         |
| ------------ | ------------ | ------------------------ | ---------------------- |
| `audit_logs` | `timestamp`  | `variables.since` (body) | `audit_logs_timestamp` |

The first incremental run reads available audit history. Later runs send the
saved timestamp through monday.com's `start_time` filter. Audit logs have no
unique event ID: use **incremental append**, accepting possible duplicates at
the checkpoint boundary or from a configured lookback. The default lookback is
zero. Verify timestamp and boundary behavior against your account before
scheduling incremental runs. See the
[audit log reference](https://developer.monday.com/api-reference/reference/audit-logs).

Items and updates remain full reads because their documented filters do not
provide a reliable saved-timestamp cursor for all edits. Use **full replace**
to remove records that the API no longer returns. `audit_logs`,
`audit_event_catalogue`, `notification_settings`, `api_usage`, and
`update_viewers` have no primary key; use full replace for their snapshot reads.

## Behavior

* **Auth**: the token is sent as a raw `Authorization` header value. Requests
  use `POST /v2` with API version `2026-07`.
* **Pagination**: numbered lists generally request 50 rows per page. Items,
  directory resources, timelines, Notetaker, and notifications use cursors.
  Audit logs use `has_more_pages` to continue through short pages. Item cursors
  expire after 60 minutes; an expired cursor fails the run.
* **Errors**: GraphQL errors fail extraction, including HTTP 200 responses
  containing partial data. Permission errors are not treated as empty resources.
* **Rate limiting**: two requests per second per connection. The engine retries
  throttled requests using `Retry-After`; monday.com also applies shared usage
  and query-complexity limits. See
  [rate limits](https://developer.monday.com/api-reference/docs/rate-limits).
* Every resource carries a `raw` remainder column preserving requested fields
  that the manifest does not project.
