> ## Documentation Index > Fetch the complete documentation index at: https://docs.markifact.com/llms.txt > Use this file to discover all available pages before exploring further. # List logs > List the operation logs your Markifact team has produced across workflows, MCP servers, and agents. Returns one log per operation your team ran, newest first, whatever produced it: a workflow node, an MCP tool call, or an agent step. API keys can only be created by team owners, so a key sees every log the team wrote, including operations run by other members and through private MCP servers. How far back you can read depends on your plan's history window. Logs outside that window are not returned even though Markifact still stores them. ## Endpoint ```http theme={"dark"} GET /v1/logs ``` ## Query Parameters | Parameter | Type | Required | Description | | ---------------- | ------- | -------- | --------------------------------------------------------------------------------------------- | | `source_type` | string | No | Only return logs from one source: `mcp`, `agent`, or `workflow`. Omit for all three. | | `status` | string | No | Only return logs with this status: `success`, `error`, `skipped`, `stop`, or `start`. | | `operation_id` | string | No | Exact operation ID, such as `gads_get_campaigns`. | | `source_id` | string | No | Only return logs from one MCP server, agent conversation, or workflow. | | `source_run_id` | string | No | Only return logs from a single workflow run. This is how you fetch every step of one run. | | `started_after` | integer | No | Only logs started at or after this Unix second. Cannot reach past your plan's history window. | | `started_before` | integer | No | Only logs started at or before this Unix second. | | `limit` | integer | No | Logs per page, 1 to 100. Defaults to 25. | | `cursor` | string | No | The `next_cursor` from the previous page. | ## Example Request ```bash theme={"dark"} curl "https://api.markifact.com/v1/logs?source_type=workflow&status=error&limit=25" \ -H "Authorization: Bearer YOUR_API_KEY" ``` ## Response ```json theme={"dark"} { "data": [ { "id": "1e38c1c2-1d85-4094-ab55-1f35c78ab436", "operation_id": "sheets_read_data", "operation_version": 1, "status": "success", "source": { "type": "workflow", "id": "bcd0e853-dd0c-4347-bd35-b86148cf2c96", "run_id": "411de6e1-9a3d-4838-a1a7-21cbc8dfe15f", "name": "Meta Ads Upload" }, "user": { "id": "Q8Wvn9gL8GUQ11SsJAnMfiyyGOA2", "email": "member@example.com" }, "started_at": 1784874165, "completed_at": 1784874172, "credits_used": 1, "cache_hit": false, "input": { "spreadsheet_id": "1AbC...", "range": "Sheet1!A1:F200" }, "output": { "rows": 199, "file_id": "f_9c21..." }, "metadata": { "node_id": "node_3", "node_label": "Read campaign sheet", "retry_attempts": 0 } } ], "has_more": true, "next_cursor": "eyJzIjoxNzg0ODc0MTY1LCJpIjoiNjI2OWRhYTUifQ" } ``` ## Pagination Read `next_cursor` from the response and pass it back as `cursor` to get the next page. When `has_more` is `false`, `next_cursor` is `null` and you have reached the end. ```bash theme={"dark"} # First page curl "https://api.markifact.com/v1/logs?limit=100" \ -H "Authorization: Bearer YOUR_API_KEY" # Next page curl "https://api.markifact.com/v1/logs?limit=100&cursor=eyJzIjoxNzg0ODc0MTY1LCJpIjoiNjI2OWRhYTUifQ" \ -H "Authorization: Bearer YOUR_API_KEY" ``` Keep every other parameter identical while paging. The cursor marks a position in the ordered result set, not the query that produced it, so changing a filter mid-walk gives you a page that does not line up with the previous one. Treat the cursor as opaque. Its contents are internal and can change. ## Log Object | Field | Type | Description | | ------------------- | --------------- | -------------------------------------------------------------------------------------------------------------------------- | | `id` | string | Log ID. | | `operation_id` | string | The Markifact operation that ran. | | `operation_version` | integer or null | Version of the operation. | | `status` | string | `success`, `error`, `skipped`, `stop`, or `start`. `start` means the operation was still running when the log was written. | | `source` | object | What ran the operation. See below. | | `user` | object | The team member who ran it, with `id` and `email`. | | `started_at` | integer | Start time in Unix seconds. | | `completed_at` | integer or null | Completion time in Unix seconds. Null while still running. | | `credits_used` | integer | Credits the operation consumed. | | `cache_hit` | boolean | Whether the result came from cache instead of the provider. | | `input` | any | Input the operation ran with. Shape depends on the operation. | | `output` | any | What the operation returned, or the error payload when `status` is `error`. Shape depends on the operation. | | `metadata` | object | Extra context. `node_id`, `node_label`, and `retry_attempts` for workflows, `tool_name` for MCP. | ### Source Object | Field | Type | Description | | -------- | -------------- | -------------------------------------------------------------------------------------------------- | | `type` | string | `mcp`, `agent`, or `workflow`. | | `id` | string or null | The MCP server, agent conversation, or workflow the operation belongs to. | | `run_id` | string or null | A single execution. Set on workflow logs, null otherwise. | | `name` | string or null | Current name of the parent. Falls back to the name recorded at log time if the parent was deleted. | Timestamps in this resource are Unix **seconds**. Connection timestamps are in milliseconds. ## Reconstructing a Workflow Run A workflow writes one log per node, all sharing a `source_run_id`. To pull a whole run in order, filter by it: ```bash theme={"dark"} curl "https://api.markifact.com/v1/logs?source_run_id=411de6e1-9a3d-4838-a1a7-21cbc8dfe15f" \ -H "Authorization: Bearer YOUR_API_KEY" ``` Logs come back newest first, so reverse them to read the run in execution order. A run that failed before any node executed, for example when the team ran out of credits, produces no logs at all. ## OpenAPI ````yaml GET /v1/logs openapi: 3.1.0 info: title: Markifact API version: 1.0.0 description: API for managing Markifact workspace connections. servers: - url: https://api.markifact.com security: - bearerAuth: [] paths: /v1/logs: get: tags: - Logs summary: List logs description: List your team's operation logs, newest first. operationId: listLogs parameters: - name: source_type in: query required: false schema: type: string enum: - mcp - agent - workflow description: Only return logs from this source. Omit for all sources. - name: status in: query required: false schema: type: string enum: - success - error - skipped - stop - start description: Only return logs with this status. - name: operation_id in: query required: false schema: type: string example: gads_get_campaigns description: Exact operation ID to filter by. - name: source_id in: query required: false schema: type: string description: >- Only return logs from one MCP server, agent conversation, or workflow. - name: source_run_id in: query required: false schema: type: string description: Only return logs from a single workflow run. - name: started_after in: query required: false schema: type: integer format: int64 minimum: 0 description: >- Only return logs started at or after this Unix second. Cannot reach past your plan's history window. - name: started_before in: query required: false schema: type: integer format: int64 minimum: 0 description: Only return logs started at or before this Unix second. - name: limit in: query required: false schema: type: integer minimum: 1 maximum: 100 default: 25 description: Logs per page. - name: cursor in: query required: false schema: type: string description: >- The next_cursor from the previous page. Keep every other filter unchanged while paging. responses: '200': description: Logs returned. content: application/json: schema: $ref: '#/components/schemas/LogPage' '400': description: Invalid cursor. content: application/json: schema: $ref: '#/components/schemas/Error' '401': description: Invalid API key. content: application/json: schema: $ref: '#/components/schemas/Error' components: schemas: LogPage: type: object properties: data: type: array items: $ref: '#/components/schemas/Log' has_more: type: boolean description: Whether more logs match beyond this page. next_cursor: type: - string - 'null' description: Pass back as `cursor` to fetch the next page. Null on the last page. required: - data - has_more Error: type: object properties: detail: {} Log: type: object properties: id: type: string description: Log ID. operation_id: type: string example: gads_get_campaigns description: Markifact operation that ran. operation_version: type: - integer - 'null' status: type: string enum: - success - error - skipped - stop - start source: $ref: '#/components/schemas/LogSource' user: $ref: '#/components/schemas/LogUser' started_at: type: integer format: int64 example: 1784874174 description: Start time in Unix seconds. completed_at: type: - integer - 'null' format: int64 description: >- Completion time in Unix seconds. Null while the operation is still running. credits_used: type: integer cache_hit: type: boolean description: Whether the result came from cache instead of the provider. input: description: Input the operation ran with. Shape depends on the operation. output: description: >- Result the operation returned, or the error payload when status is error. Shape depends on the operation. metadata: type: object additionalProperties: true description: >- Extra context. Keys depend on the source: node_id, node_label, and retry_attempts for workflows, tool_name for MCP. required: - id - operation_id - status - source - user - started_at - credits_used - cache_hit - metadata LogSource: type: object properties: type: type: string enum: - mcp - agent - workflow description: What ran the operation. id: type: - string - 'null' description: 'Parent object ID: the MCP server, agent conversation, or workflow.' run_id: type: - string - 'null' description: Single execution ID. Set on workflow logs, null otherwise. name: type: - string - 'null' description: >- Current parent name, falling back to the name captured at log time if the parent was deleted. required: - type LogUser: type: object properties: id: type: - string - 'null' description: Markifact user ID. email: type: - string - 'null' description: Email of the team member who ran the operation. securitySchemes: bearerAuth: type: http scheme: bearer ````