Skip to main content
GET /workspaces/usage Returns agent execution history for a workspace. Results combine current executions, archived rows, and edited chats (see Item fields). Responses are sorted by created_at descending, then id descending. Your API token must be allowed to use agents (use_agents), and you must have access to the workspace—the same layers as other agent and file API routes (Sanctum authentication, workspace access check).

​
Code Examples

curl --request GET \
  --url 'https://api.tess.im/workspaces/usage?range=7d&page=1&per_page=20' \
  --header 'Authorization: Bearer YOUR_API_KEY' \
  --header 'Accept: application/json' \
  --header 'x-workspace-id: YOUR_WORKSPACE_ID'

​
Headers

ParameterTypeRequiredDescription
AuthorizationstringYesBearer token (API key). The token must include the use_agents ability.
x-workspace-idintegerYesNumeric workspace ID. Invalid or missing values produce a 422 error. You must belong to this workspace or the API returns 403.
AcceptstringNoUse application/json if your client needs an explicit JSON accept header.

​
Query parameters

ParameterTypeRequiredDescriptionDefault / rules
rangestringNoRelative window: 1d, 7d, or 30d. Ignored if both start_date and end_date are sent.If no date parameters are sent, the effective default is 30d.
start_datedate (YYYY-MM-DD)ConditionalStart of a custom range (inclusive).Required with end_date.
end_datedate (YYYY-MM-DD)ConditionalEnd of a custom range (inclusive). Must be greater than or equal to start_date.Required with start_date. The span cannot exceed 90 days; otherwise 422.
user_idintegerNoFilter by the user who ran the agent.Without workspace permission to read other users’ activity, you only see your own executions; filtering by another user’s id returns 403.
typestringNoAgent type filter: all, chat, image, text, voiceover, video, code.all or omitted means no type filter.
pageintegerNoPage number.Default 1, minimum 1.
per_pageintegerNoPage size.Default 20, between 1 and 100.
Date windows
  • With start_date + end_date: the inclusive range is capped at 90 days.
  • With range: the window is relative to the end of the current day (1d last 24 hours from that instant; 7d / 30d last seven or thirty calendar days from that end).
  • If the usage_history_min_date feature is enabled in app settings, the effective start of the range is not earlier than that date (silent clamp).
Caching
  • Listings are cached about 60 seconds per workspace, filters, and page. Identical requests within that window may return the same payload.

​
Response

{
  "items": [
    {
      "id": "183450",
      "created_at": "2026-04-07 16:17:09",
      "user_id": 16643,
      "type": "chat",
      "status": "succeeded",
      "email": "user@example.com",
      "credits": 1.5,
      "name": "My personal assistant",
      "slug": "9b4994e3-07e9-4163-b5c0-9c4f18945eda-my-personal-assistant",
      "output": "This content is only available on Tess",
      "root_id": null,
      "execution_origin": "Platform",
      "source": "current",
      "used_model": "gpt-4o-mini",
      "tokens": {
        "input": 1240,
        "output": 320,
        "total": 1560
      },
      "link": "/dashboard/user/ai/chat/ai-chat/9b4994e3-07e9-4163-b5c0-9c4f18945eda-my-personal-assistant?_chat_id=183450"
    }
  ],
  "pagination": {
    "current_page": 1,
    "per_page": 20,
    "has_more": false
  }
}
Pagination uses has_more: the service requests per_page + 1 rows; if the extra row exists, has_more is true and only the first per_page items appear in items.

​
Item fields

FieldDescription
idExecution id, or a synthetic id such as chat-edited-{user_openai_id} for edited chats.
created_atTimestamp of the execution.
user_idUser who ran the agent.
typeAgent type from the agent definition (for example chat, image, text, voiceover, video, code).
statussucceeded or failed (non-success statuses map to failed).
emailEmail of the executing user.
creditsCredits charged; 0 when status is not succeeded.
nameAgent title.
slugAgent slug.
outputFor current rows: for image, video, and voiceover this is the real output; for other types it is the fixed text This content is only available on Tess. For archived: This item was deleted. For edited: This item was edited.
root_idConversation root for chats, when present.
execution_originTaken from execution metadata when available.
sourcecurrent, archived, or edited.
used_modelModel name when stored (user_openai.used_model or archived JSON); null for edited rows in the current implementation.
tokensObject { "input", "output", "total" } when detailed_credits indicates token billing; otherwise null (common for image, video, voiceover).
linkURL or app path to open the resource when applicable; null for archived and edited, and for types without a link. Chat (current): path /dashboard/user/ai/chat/ai-chat/{slug}?_chat_id={id} using root_id when set, else row id. Image / video / voiceover (current): signed or public URL from storage when output is usable; otherwise null.

​
Errors

StatusWhen
401Missing or invalid authentication (Sanctum).
403No access to the workspace, or user_id targets another user without permission to view their executions.
422Invalid query parameters (Laravel validation), invalid or missing workspace id, invalid dates, or a custom range longer than 90 days.
Validation errors use the standard Laravel error payload; some workspace errors return JSON { "message": "..." } with a translated message.