openapi: 3.2.0 info: description: Relevance AI Endpoints title: Relevance AI Endpoints Analytics API version: latest contact: email: dev@tryrelevance.com servers: - url: /latest security: - AuthorizationHeader: - Authorization tags: - name: Analytics paths: /analytics/get_metrics: post: summary: GetMetrics tags: - Analytics description: 'Get metrics data with time-series points and summary statistics ### Required permissions > []' operationId: GetMetrics parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetMetricsInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetMetricsOutput' /analytics/get_agent_breakdown: post: summary: GetAgentBreakdown tags: - Analytics description: 'Get agent breakdown data with task counts, error rates, and top actions ### Required permissions > []' operationId: GetAgentBreakdown parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetAgentBreakdownInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetAgentBreakdownOutput' /analytics/get_action_breakdown: post: summary: GetActionBreakdown tags: - Analytics description: 'Get action breakdown data with execution counts, credits used, and top agents ### Required permissions > []' operationId: GetActionBreakdown parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetActionBreakdownInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetActionBreakdownOutput' /analytics/get_workforce_breakdown: post: summary: GetWorkforceBreakdown tags: - Analytics description: 'Get workforce breakdown data with task counts, error rates, credits used, and top agents ### Required permissions > []' operationId: GetWorkforceBreakdown parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetWorkforceBreakdownInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetWorkforceBreakdownOutput' /analytics/get_agent_llm_usage_by_model: post: summary: GetAgentModelBreakdown tags: - Analytics description: 'Get agent usage breakdown by LLM model with credits and input/output token sums ### Required permissions > []' operationId: GetAgentModelBreakdown parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetAgentModelBreakdownInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetAgentModelBreakdownOutput' /analytics/get_agent_errors: post: summary: GetAgentErrors tags: - Analytics description: 'Get error messages from agents and tool runs within a time range ### Required permissions > []' operationId: GetAgentErrors parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetAgentErrorsInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetAgentErrorsOutput' /analytics/get_error_time_series: get: summary: GetErrorTimeSeries tags: - Analytics description: 'Get detailed error records with full execution context ### Required permissions > []' operationId: GetErrorTimeSeries parameters: - in: query name: start_date schema: type: string format: date-time description: '' - in: query name: end_date schema: type: string format: date-time description: '' - in: query name: agent_ids schema: type: array items: type: string description: '' - in: query name: workforce_ids schema: type: array items: type: string description: '' - in: query name: source_filter schema: type: string enum: - all - workforce - agent - tool description: 'Filter errors by source type: all (no filter), workforce (only errors from workforce tasks), agent (agent conversation errors), tool (tool execution errors)' description: 'Filter errors by source type: all (no filter), workforce (only errors from workforce tasks), agent (agent conversation errors), tool (tool execution errors)' - in: query name: page_size schema: type: number minimum: 1 maximum: 100 default: 20 description: '' - in: query name: cursor schema: type: string description: '' - in: query name: timezone schema: type: string minLength: 1 maxLength: 64 description: IANA timezone identifier (e.g., 'America/New_York', 'Europe/London'). Used for date aggregation in charts. description: IANA timezone identifier (e.g., 'America/New_York', 'Europe/London'). Used for date aggregation in charts. responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetErrorTimeSeriesOutput' /analytics/get_task_timeline: get: summary: GetTaskTimeline tags: - Analytics description: 'Get day-bucketed counts for one of the Tasks page tab timelines — `view` picks Approvals, Escalated, or Errored. Auth matches ListConversations so any project member viewing the tab can also see the chart. ### Required permissions > [ { "actions": [ "agents:read", "knowledge_sets:read" ] } ]' operationId: GetTaskTimeline parameters: - in: query name: view schema: type: string enum: - approvals - escalated - errored description: 'Which tab''s timeline to fetch counts for: ''approvals'' (pending-approval conversations — the Approvals tab), ''escalated'' (conversations currently in the escalated state), or ''errored'' (conversations in an errored state with active failures).' description: 'Which tab''s timeline to fetch counts for: ''approvals'' (pending-approval conversations — the Approvals tab), ''escalated'' (conversations currently in the escalated state), or ''errored'' (conversations in an errored state with active failures).' - in: query name: start_date schema: type: string format: date-time description: Inclusive start of the time range, UTC ISO 8601 format (e.g., '2024-01-15T09:00:00Z'). description: Inclusive start of the time range, UTC ISO 8601 format (e.g., '2024-01-15T09:00:00Z'). - in: query name: end_date schema: type: string format: date-time description: Inclusive end of the time range, UTC ISO 8601 format (e.g., '2024-01-15T09:00:00Z'). description: Inclusive end of the time range, UTC ISO 8601 format (e.g., '2024-01-15T09:00:00Z'). - in: query name: agent_ids schema: type: array items: type: string maxLength: 128 maxItems: 100 description: '' - in: query name: timezone schema: type: string maxLength: 64 description: IANA timezone identifier (e.g., 'America/New_York', 'Europe/London'). Used for date aggregation in charts. description: IANA timezone identifier (e.g., 'America/New_York', 'Europe/London'). Used for date aggregation in charts. responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetTaskTimelineOutput' /organizations/{organization_id}/analytics/concurrency_time_series: get: summary: GetConcurrencyTimeSeriesByOrgId tags: - Analytics description: 'Get concurrency usage time-series data for an organization (canonical org-scoped path) ### Required permissions > []' operationId: GetConcurrencyTimeSeriesByOrgId parameters: - name: organization_id in: path description: ID of organization required: true schema: type: string maxLength: 240 pattern: ^[a-zd._-]+$ - in: query name: start_date schema: type: string format: date-time description: Start date for the query (ISO 8601 format) description: Start date for the query (ISO 8601 format) - in: query name: end_date schema: type: string format: date-time description: End date for the query (ISO 8601 format) description: End date for the query (ISO 8601 format) - in: query name: granularity schema: type: string enum: - 15m - 4h - 1d description: Time bucket granularity description: Time bucket granularity - in: query name: aggregation schema: type: string enum: - max - avg default: max description: Aggregation function applied within each bucket. Defaults to max description: Aggregation function applied within each bucket. Defaults to max - in: query name: project_id schema: type: string description: Optional project ID to filter by. Not yet supported — returns 400 if set description: Optional project ID to filter by. Not yet supported — returns 400 if set - in: query name: group_by schema: type: string enum: - project description: Split the concurrency series per project when set to "project". description: Split the concurrency series per project when set to "project". - in: query name: detail_scope schema: type: string enum: - organization - current_project description: Requested detail scope for the per-project breakdown. Org admins receive full per-project detail; other callers only receive full detail for detail_project_id, with every other project anonymised into a single bucket. The server clamps this to the caller's role, so it can only narrow what is returned. description: Requested detail scope for the per-project breakdown. Org admins receive full per-project detail; other callers only receive full detail for detail_project_id, with every other project anonymised into a single bucket. The server clamps this to the caller's role, so it can only narrow what is returned. - in: query name: detail_project_id schema: type: string minLength: 1 description: Project display ID that can receive full detail in the per-project breakdown when detail_scope is current_project. description: Project display ID that can receive full detail in the per-project breakdown when detail_scope is current_project. - in: query name: timezone schema: type: string description: IANA timezone identifier (e.g., 'America/New_York', 'Europe/London'). Used for date aggregation in charts. description: IANA timezone identifier (e.g., 'America/New_York', 'Europe/London'). Used for date aggregation in charts. responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetConcurrencyTimeSeriesByOrgIdOutput' /organizations/{organization_id}/analytics/live_jobs: get: summary: GetLiveJobsByOrgId tags: - Analytics description: 'Get live scheduler-backed running and queued jobs for an organization ### Required permissions > []' operationId: GetLiveJobsByOrgId parameters: - name: organization_id in: path description: ID of organization required: true schema: type: string maxLength: 240 pattern: ^[a-zd._-]+$ - in: query name: cursor schema: type: string description: Opaque pagination cursor from the previous response description: Opaque pagination cursor from the previous response - in: query name: limit schema: type: integer minimum: 1 maximum: 500 description: Maximum live jobs to return. Defaults to 100 description: Maximum live jobs to return. Defaults to 100 - in: query name: state schema: type: string enum: - running - queued description: Restrict the returned page to one state. Counts in metadata stay organization-wide and unfiltered. Omitted returns both, interleaved in dispatch order description: Restrict the returned page to one state. Counts in metadata stay organization-wide and unfiltered. Omitted returns both, interleaved in dispatch order - in: query name: priority_tier schema: type: string enum: - high - medium - low - untiered description: Restrict the returned page to one priority tier, 'untiered' included. Counts in metadata stay organization-wide and unfiltered, as with state. Omitted returns every tier description: Restrict the returned page to one priority tier, 'untiered' included. Counts in metadata stay organization-wide and unfiltered, as with state. Omitted returns every tier - in: query name: detail_scope schema: type: string enum: - organization - current_project description: Whether to return organization-wide full detail or only full detail for the current project description: Whether to return organization-wide full detail or only full detail for the current project - in: query name: detail_project_id schema: type: string description: Project display ID that can receive full row detail when detail_scope is current_project description: Project display ID that can receive full row detail when detail_scope is current_project responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetLiveJobsByOrgIdOutput' /task_ops/tab_counts: get: summary: GetTaskOpsTabCounts tags: - Analytics description: 'Get the built-in TaskOps tab counts (Approvals, Escalated, Errors, All tasks) in a single response. ### Required permissions > [ { "actions": [ "agents:read", "knowledge_sets:read" ] } ]' operationId: GetTaskOpsTabCounts parameters: [] responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetTaskOpsTabCountsOutput' /analytics/get_usage: get: summary: GetUsage tags: - Analytics description: 'Get usage data from agents and tool runs within a time range ### Required permissions > []' operationId: GetUsage parameters: - in: query name: group schema: type: string enum: - user - agent - tool - workforce description: The actor associated with the usage description: The actor associated with the usage - in: query name: metric schema: type: string enum: - credit - action description: The metric to be measured description: The metric to be measured - in: query name: from schema: type: string format: date description: '' - in: query name: to schema: type: string format: date description: '' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetUsageOutput' /analytics/tool/breakdown: post: summary: GetToolUsageBreakdown tags: - Analytics description: 'Per-tool totals across the date range. Returns `{ meta, rows, totals.grand }` — one row per tool (entity_id + metadata + runs/actions/credits/credits_per_run/error_rate) ordered by runs DESC. ### Required permissions > []' operationId: GetToolUsageBreakdown parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetToolUsageBreakdownInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetToolUsageBreakdownOutput' /analytics/tool/breakdown/timeseries: post: summary: GetToolUsageBreakdownTimeseries tags: - Analytics description: 'Per-(tool, day) usage zero-filled across the date range. Returns `{ meta, entities, series, totals }` — `entities` carries per-tool metadata + totals (uncapped), `series` is long-format `(entity_id, date, ...)` chart rows capped at the top 1000 tools by run volume, `totals.by_date` is per-day rollups across the capped `series` only. For projects with >1000 tools the caps diverge: `totals.grand` covers all entities while `totals.by_date` and `series` cover only the top 1000, so `sum(totals.by_date.runs) []' operationId: GetToolUsageBreakdownTimeseries parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetToolUsageBreakdownTimeseriesInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetToolUsageBreakdownTimeseriesOutput' /analytics/tool/timeseries: post: summary: GetToolUsageTimeseries tags: - Analytics description: 'Per-day usage across selected tools (default = all), zero-filled across the date range. Returns `{ meta, rows, totals.grand }` — `rows` ordered by date ASC, one per day with combined runs/actions/credits/error_rate. ### Required permissions > []' operationId: GetToolUsageTimeseries parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetToolUsageTimeseriesInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetToolUsageTimeseriesOutput' /analytics/agent/breakdown: post: summary: GetAgentUsageBreakdown tags: - Analytics description: 'Per-agent totals across the date range. Returns `{ meta, rows, totals.grand }` — one row per agent (entity_id + metadata + tasks/actions/credits/credits_per_task/error_rate) ordered by tasks DESC. When `children=yes`, each row carries a `children` array of per-tool rollups stitched at the agent grain (tool-grained children remain runs-keyed since primary metrics are per-entity). ### Required permissions > []' operationId: GetAgentUsageBreakdown parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetAgentUsageBreakdownInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetAgentUsageBreakdownOutput' /analytics/agent/breakdown/timeseries: post: summary: GetAgentUsageBreakdownTimeseries tags: - Analytics description: 'Per-(agent, day) usage zero-filled across the date range. Returns `{ meta, entities, series, totals }` — `entities` carries per-agent metadata + totals (uncapped), `series` is long-format `(entity_id, date, ...)` chart rows capped at the top 1000 agents by task volume, `totals.by_date` is per-day rollups across the capped `series` only. For projects with >1000 agents the caps diverge: `totals.grand` covers all entities while `totals.by_date` and `series` cover only the top 1000, so `sum(totals.by_date.tasks) []' operationId: GetAgentUsageBreakdownTimeseries parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetAgentUsageBreakdownTimeseriesInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetAgentUsageBreakdownTimeseriesOutput' /analytics/agent/timeseries: post: summary: GetAgentUsageTimeseries tags: - Analytics description: 'Per-day agent usage across selected agents (default = all), zero-filled across the date range. Returns `{ meta, rows, totals.grand }` — `rows` ordered by date ASC, one per day with combined tasks/actions/agent_model_credits/tool_credits/error_rate (sum the two credit columns client-side for a full credit total). ### Required permissions > []' operationId: GetAgentUsageTimeseries parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetAgentUsageTimeseriesInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetAgentUsageTimeseriesOutput' /analytics/workforce/breakdown: post: summary: GetWorkforceUsageBreakdown tags: - Analytics description: 'Per-workforce totals across the date range. Returns `{ meta, rows, totals.grand }` — one row per workforce (entity_id + metadata + tasks/actions/credits/credits_per_task/error_rate) ordered by tasks DESC. `error_rate` counts only `state_completed` as success (state_idle does not apply to workforces, unlike agents). When `children=yes`, each row carries a `children` array of per-agent rollups (each with its own per-tool `children` array) stitched at the workforce grain — the full workforce -> agent -> tool tree. Agent-grained children remain tasks-keyed and tool-grained children remain runs-keyed (primary metrics are per-entity). ### Required permissions > []' operationId: GetWorkforceUsageBreakdown parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetWorkforceUsageBreakdownInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetWorkforceUsageBreakdownOutput' /analytics/workforce/breakdown/timeseries: post: summary: GetWorkforceUsageBreakdownTimeseries tags: - Analytics description: 'Per-(workforce, day) usage zero-filled across the date range. Returns `{ meta, entities, series, totals }` — `entities` carries per-workforce metadata + totals (uncapped), `series` is long-format `(entity_id, date, ...)` chart rows capped at the top 1000 workforces by task volume, `totals.by_date` is per-day rollups across the capped `series` only. For projects with >1000 workforces the caps diverge: `totals.grand` covers all entities while `totals.by_date` and `series` cover only the top 1000, so `sum(totals.by_date.tasks) []' operationId: GetWorkforceUsageBreakdownTimeseries parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetWorkforceUsageBreakdownTimeseriesInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetWorkforceUsageBreakdownTimeseriesOutput' /analytics/workforce/timeseries: post: summary: GetWorkforceUsageTimeseries tags: - Analytics description: 'Per-day workforce usage across selected workforces (default = all), zero-filled across the date range. Returns `{ meta, rows, totals.grand }` — `rows` ordered by date ASC, one per day with combined tasks/actions/agent_model_credits/tool_credits/error_rate (sum the two credit columns client-side for a full credit total). ### Required permissions > []' operationId: GetWorkforceUsageTimeseries parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetWorkforceUsageTimeseriesInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetWorkforceUsageTimeseriesOutput' /analytics/project/timeseries: post: summary: GetProjectUsageTimeseries tags: - Analytics description: 'Per-day project-wide usage, zero-filled across the date range. Returns `{ meta, rows, totals.grand }` with agent_tasks, workforce_tasks, tool_runs, actions, and two credit columns: `tool_credits` (tool execution incl. AI model calls made inside tools) and `agent_model_credits` (agents'' own AI model calls). Workforce credits are excluded to avoid double-counting; sum the two columns client-side for a full credit total. ### Required permissions > []' operationId: GetProjectUsageTimeseries parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetProjectUsageTimeseriesInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetProjectUsageTimeseriesOutput' /analytics/model/breakdown: post: summary: GetModelUsageBreakdown tags: - Analytics description: 'Per-(model, context) AI model usage across the date range, where `context` labels where each call occurred in the workforce/agent/tool tree (six buckets: tool, agent_model, agent_tool, workforce_tool, workforce_agent_tool, workforce_agent_model). Returns `{ meta, rows, totals }` — `rows` ordered by credits DESC, model ASC, context ASC; `totals.by_context` collapses model out for per-context subtotals (contexts with zero usage are omitted); `totals.grand` sums across the whole window. Filter via `model_ids` and/or `contexts`. ### Required permissions > []' operationId: GetModelUsageBreakdown parameters: [] requestBody: content: application/json: schema: $ref: '#/components/schemas/GetModelUsageBreakdownInput' responses: '200': description: successful operation content: application/json: schema: $ref: '#/components/schemas/GetModelUsageBreakdownOutput' components: schemas: GetAgentErrorsOutput: type: object properties: errors: type: array items: type: object properties: error_message: type: string description: Human-readable error message timestamp: description: When the error occurred (ISO 8601 format) agent_id: type: - string - 'null' description: Agent display ID associated with the error tool_id: type: - string - 'null' description: Tool display ID (only for tool execution errors) conversation_id: type: string description: Conversation display ID where the error occurred executor_type: type: - string - 'null' description: Type of executor that triggered the error (only for tool errors) step_name: type: - string - 'null' description: Name of the step where the error occurred (only for tool errors) raw_error: type: - string - 'null' description: Raw error details if available error_source: type: string enum: - tool - agent description: Source of the error - tool execution or agent level state_metadata: type: - object - 'null' description: Additional error context from conversation state metadata (agent errors only) required: - error_message - timestamp - agent_id - conversation_id - error_source additionalProperties: false description: List of error items sorted by timestamp descending next_cursor: type: object properties: timestamp: type: string description: ISO 8601 timestamp of the last error returned required: - timestamp additionalProperties: false has_more: type: boolean description: Whether there are more results after this page required: - errors - has_more additionalProperties: false GetUsageOutput: type: object properties: query: type: object properties: group: type: string enum: - user - agent - tool - workforce description: The actor associated with the usage metric: type: string enum: - credit - action description: The metric to be measured from: type: string format: date to: type: string format: date required: - metric - group - from - to additionalProperties: false series: type: array items: type: object properties: id: type: string actor: type: object properties: id: type: string type: type: string enum: - user - agent - tool - workforce description: The actor associated with the usage machine: type: boolean metadata: type: object properties: label: type: string avatar_url: type: string additionalProperties: false required: - type - id additionalProperties: false metric: type: string enum: - credit - action description: The metric to be measured value: type: number series: type: array items: type: object properties: id: type: string actor: type: object properties: id: type: string type: type: string enum: - user - agent - tool - workforce description: The actor associated with the usage machine: type: boolean metadata: type: object properties: label: type: string avatar_url: type: string additionalProperties: false required: - type - id additionalProperties: false metric: type: string enum: - credit - action description: The metric to be measured value: type: number required: - id - actor - metric - value additionalProperties: false required: - id - actor - metric - value description: A time series representing usage data over time additionalProperties: false required: - query - series additionalProperties: false GetModelUsageBreakdownOutput: type: object properties: meta: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start of the data range covered by this response (as specified by the request) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end of the data range covered by this response (as specified by the request) interval: type: string enum: - day - total description: Bucket granularity. `day` on time-bucketed endpoints (timeseries / breakdown_timeseries); `total` on plain breakdown endpoints (single bucket covering the whole window). Always present. timezone: type: string description: IANA timezone the daily buckets were materialized in (mirrors the org's internally set timezone). entity: type: string enum: - tool - agent - workforce - project - model description: Entity scope of the response. Each v2 analytics endpoint owns a single entity type; `project` denotes project-wide rollups; `model` denotes AI model usage rollups. required: - start_date - end_date - interval - timezone - entity additionalProperties: false rows: type: array items: type: object properties: model: type: string description: AI model identifier (e.g. 'gpt-4o', 'claude-3-7-sonnet-latest'). model_name: type: - string - 'null' description: Display label for the model (e.g. 'GPT-4o'). Null for unknown model ids — the frontend should render a fallback. description: type: - string - 'null' description: Human-readable model description. Null for unknown model ids. group_name: type: - string - 'null' description: Vendor / group label (e.g. 'OpenAI', 'Anthropic'). Null for unknown model ids. is_shutdown: type: boolean description: True if the provider has shut the model down. False for unknown model ids. icon: type: - string - 'null' description: CDN URL for the model vendor's icon (SVG). Null when the model has no known parent vendor. Matches the `icon` field used by the v2 Agent / Tool / Workforce breakdown endpoints. runs: type: number description: Count of AI model call invocations. input_tokens: type: number description: Total prompt / input tokens consumed. output_tokens: type: number description: Total completion / output tokens produced. credits: type: number description: Total credits consumed. children: type: array items: type: object properties: context: type: string enum: - tool - agent_model - agent_tool - workforce_tool - workforce_agent_tool - workforce_agent_model description: Where the AI model call happened in the workforce/agent/tool tree. `tool` = standalone tool run, no agent or workforce; `agent_model` = an agent's own AI model call, no tool or workforce; `agent_tool` = a tool inside an agent task, no workforce; `workforce_tool` = a workforce-driven tool with no agent; `workforce_agent_tool` = a tool inside an agent task inside a workforce; `workforce_agent_model` = an agent's own AI model call inside a workforce, no tool. runs: type: number description: Count of AI model call invocations. input_tokens: type: number description: Total prompt / input tokens consumed. output_tokens: type: number description: Total completion / output tokens produced. credits: type: number description: Total credits consumed. required: - context - runs - input_tokens - output_tokens - credits additionalProperties: false description: Per-context rollups, one entry per context bucket observed for this model in the window (ordered credits DESC, context ASC). Parent-level metrics (runs / input_tokens / output_tokens / credits) are sums across `children`. Always present, never empty. required: - model - model_name - description - group_name - is_shutdown - icon - runs - input_tokens - output_tokens - credits - children additionalProperties: false totals: type: object properties: by_context: type: array items: type: object properties: context: type: string enum: - tool - agent_model - agent_tool - workforce_tool - workforce_agent_tool - workforce_agent_model description: Where the AI model call happened in the workforce/agent/tool tree. `tool` = standalone tool run, no agent or workforce; `agent_model` = an agent's own AI model call, no tool or workforce; `agent_tool` = a tool inside an agent task, no workforce; `workforce_tool` = a workforce-driven tool with no agent; `workforce_agent_tool` = a tool inside an agent task inside a workforce; `workforce_agent_model` = an agent's own AI model call inside a workforce, no tool. runs: type: number description: Count of AI model call invocations. input_tokens: type: number description: Total prompt / input tokens consumed. output_tokens: type: number description: Total completion / output tokens produced. credits: type: number description: Total credits consumed. required: - context - runs - input_tokens - output_tokens - credits additionalProperties: false description: Subtotals collapsed across model, one per context bucket that had any usage in the window. Contexts with zero usage are omitted. grand: type: object properties: runs: type: number description: Count of AI model call invocations. input_tokens: type: number description: Total prompt / input tokens consumed. output_tokens: type: number description: Total completion / output tokens produced. credits: type: number description: Total credits consumed. required: - runs - input_tokens - output_tokens - credits additionalProperties: false required: - by_context - grand additionalProperties: false required: - meta - rows - totals additionalProperties: false GetAgentUsageTimeseriesOutput: type: object properties: meta: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start of the data range covered by this response (as specified by the request) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end of the data range covered by this response (as specified by the request) interval: type: string enum: - day - total description: Bucket granularity. `day` on time-bucketed endpoints (timeseries / breakdown_timeseries); `total` on plain breakdown endpoints (single bucket covering the whole window). Always present. timezone: type: string description: IANA timezone the daily buckets were materialized in (mirrors the org's internally set timezone). entity: type: string enum: - tool - agent - workforce - project - model description: Entity scope of the response. Each v2 analytics endpoint owns a single entity type; `project` denotes project-wide rollups; `model` denotes AI model usage rollups. required: - start_date - end_date - interval - timezone - entity additionalProperties: false rows: type: array items: type: object properties: date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Bucket date tasks: type: number description: Count of tasks across all selected agents actions: type: number description: Count of actions across all selected agents agent_model_credits: type: number description: Credits consumed by the agent's own model calls for the bucket. Pairs with tool_credits. tool_credits: type: number description: Credits consumed by tools the agent ran for the bucket. Pairs with agent_model_credits. error_rate: type: number description: Fraction of tasks not in a success state, in [0, 1]; 0 when there are no tasks in the bucket required: - date - tasks - actions - agent_model_credits - tool_credits - error_rate additionalProperties: false totals: type: object properties: grand: type: object properties: tasks: type: number description: Count of tasks (the agent orchestration unit; agent primary metric) actions: type: number description: Count of actions agent_model_credits: type: number description: Credits consumed by agents' own model calls. Pairs with tool_credits. tool_credits: type: number description: Credits consumed by tools agents ran. Pairs with agent_model_credits. credits_per_task: type: number description: (agent_model_credits + tool_credits) / tasks ratio. 0 when tasks is 0. error_rate: type: number description: Fraction of tasks not in a success state (success = state_completed + state_idle), in [0, 1]; 0 when tasks is 0 required: - tasks - actions - agent_model_credits - tool_credits - credits_per_task - error_rate additionalProperties: false required: - grand additionalProperties: false required: - meta - rows - totals additionalProperties: false GetWorkforceUsageTimeseriesInput: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start date (YYYY-MM-DD) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end date (YYYY-MM-DD) workforce_ids: type: array items: type: string minLength: 1 maxLength: 256 maxItems: 1000 description: Optional list of workforce display IDs to filter by. Duplicates are ignored. Maximum 1000 IDs. required: - start_date - end_date additionalProperties: false GetWorkforceUsageBreakdownTimeseriesOutput: type: object properties: meta: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start of the data range covered by this response (as specified by the request) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end of the data range covered by this response (as specified by the request) interval: type: string enum: - day - total description: Bucket granularity. `day` on time-bucketed endpoints (timeseries / breakdown_timeseries); `total` on plain breakdown endpoints (single bucket covering the whole window). Always present. timezone: type: string description: IANA timezone the daily buckets were materialized in (mirrors the org's internally set timezone). entity: type: string enum: - tool - agent - workforce - project - model description: Entity scope of the response. Each v2 analytics endpoint owns a single entity type; `project` denotes project-wide rollups; `model` denotes AI model usage rollups. required: - start_date - end_date - interval - timezone - entity additionalProperties: false entities: type: array items: type: object properties: entity_id: type: string description: Workforce display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' tasks: type: number description: Count of tasks (the workforce orchestration unit; workforce primary metric) error_rate: type: number description: Fraction of tasks not completed (workforce success = state_completed only; state_idle does not apply to workforces), in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_task: type: number description: credits / tasks ratio. 0 when tasks is 0. children: type: array items: type: object properties: entity_id: type: string description: Agent display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' tasks: type: number description: Count of tasks (the agent orchestration unit; agent primary metric) error_rate: type: number description: Fraction of tasks not in a success state (agent success = state_completed + state_idle), in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_task: type: number description: credits / tasks ratio. 0 when tasks is 0. children: type: array items: type: object properties: entity_id: type: string description: Tool display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' runs: type: number description: Count of runs error_rate: type: number description: Fraction of tool runs that did not complete, in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_run: type: number description: credits / runs ratio. 0 when runs is 0. required: - entity_id - name - description - icon - runs - error_rate - actions - credits - credits_per_run additionalProperties: false description: Present only when the request `children=yes`. Per-tool rollups stitched at the agent grain — one entry per tool the agent invoked within this workforce over the window. Tool-grained children remain runs-keyed (primary metric per entity). required: - entity_id - name - description - icon - tasks - error_rate - actions - credits - credits_per_task additionalProperties: false description: Present only when the request `children=yes`. Per-agent rollups stitched at the workforce grain. Each agent child may carry its own nested `children` array of per-tool rollups (the full workforce -> agent -> tool tree). required: - entity_id - name - description - icon - tasks - error_rate - actions - credits - credits_per_task additionalProperties: false series: type: array items: type: object properties: entity_id: type: string description: Workforce display ID; joins to `entities[].entity_id` date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Bucket date tasks: type: number description: Count of tasks for this (workforce, day) credits: type: number description: Total credits for the day actions: type: number description: Count of actions error_rate: type: number description: Fraction of tasks not completed (success = state_completed), in [0, 1]; 0 when there are no tasks required: - entity_id - date - tasks - credits - actions - error_rate additionalProperties: false totals: type: object properties: grand: type: object properties: tasks: type: number description: Count of tasks (the workforce orchestration unit; workforce primary metric) actions: type: number description: Count of actions agent_model_credits: type: number description: Credits consumed by agents' own model calls inside the selected workforces. Pairs with tool_credits (disjoint; mirrors project-level semantics so workforce/agent/tool credits aren't double-counted). tool_credits: type: number description: Credits consumed by tools run inside the selected workforces. Pairs with agent_model_credits. credits_per_task: type: number description: (agent_model_credits + tool_credits) / tasks ratio. 0 when tasks is 0. error_rate: type: number description: Fraction of tasks not completed (workforce success = state_completed only; no state_idle exists at the workforce level), in [0, 1]; 0 when tasks is 0 required: - tasks - actions - agent_model_credits - tool_credits - credits_per_task - error_rate additionalProperties: false by_date: type: array items: type: object properties: date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Bucket date tasks: type: number description: Count of tasks across all selected workforces actions: type: number description: Count of actions across all selected workforces credits: type: number description: Total credits consumed for the bucket error_rate: type: number description: Fraction of tasks not completed (success = state_completed), in [0, 1]; 0 when there are no tasks in the bucket required: - date - tasks - actions - credits - error_rate additionalProperties: false required: - grand - by_date additionalProperties: false required: - meta - entities - series - totals additionalProperties: false GetAgentBreakdownOutput: type: object properties: agents: type: array items: type: object properties: agent_id: type: string description: Agent display ID agent_name: type: string description: Agent name (fallback to default if null) task_count: type: number description: Total number of tasks for this agent error_rate: type: number description: Error rate percentage (0-100) error_count: type: number description: Number of errored tasks (the raw count behind error_rate). total_actions: type: number description: Total number of actions executed by this agent credits_used: type: number description: Total credits used by this agent top_actions: type: array items: type: object properties: action_name: type: string description: Name of the action/tool execution_count: type: number description: Number of times this action was executed required: - action_name - execution_count additionalProperties: false description: Top 3 most used actions by this agent maxItems: 3 required: - agent_id - agent_name - task_count - error_rate - total_actions - credits_used - top_actions additionalProperties: false description: Agent breakdown data required: - agents additionalProperties: false GetWorkforceBreakdownOutput: type: object properties: workforces: type: array items: type: object properties: workforce_id: type: string description: Workforce display ID workforce_name: type: string description: Workforce name (fallback to default if null) workforce_tasks: type: number description: Number of unique workforce tasks (count distinct workforce_task_display_id) agent_tasks: type: number description: Total number of agent tasks/conversations spawned by the workforce error_rate: type: number description: Error rate percentage (0-100) based on agent tasks error_count: type: number description: Number of errored agent tasks (the raw count behind error_rate). total_actions: type: number description: Total number of actions executed by this workforce credits_used: type: number description: Total credits used by this workforce top_agents: type: array items: type: object properties: agent_id: type: string description: Agent display ID agent_name: type: string description: Agent name (fallback to default if null) task_count: type: number description: Number of tasks this agent executed using this action required: - agent_id - agent_name - task_count additionalProperties: false description: Top 3 agents used by this workforce maxItems: 3 required: - workforce_id - workforce_name - workforce_tasks - agent_tasks - error_rate - total_actions - credits_used - top_agents additionalProperties: false description: Workforce breakdown data required: - workforces additionalProperties: false GetTaskOpsTabCountsOutput: type: object properties: approvals: type: integer minimum: 0 description: Count of conversations currently in the pending-approval state. escalated: type: integer minimum: 0 description: Count of conversations currently in the escalated state. errored: type: integer minimum: 0 description: Count of conversations in one of the errored states or with at least one active failure. all_tasks: type: integer minimum: 0 description: Count of non-completed conversations, matching the All tasks tab. required: - approvals - escalated - errored - all_tasks additionalProperties: false GetWorkforceBreakdownInput: type: object properties: start_date: description: Start date for the workforce breakdown query (ISO 8601 format) end_date: description: End date for the workforce breakdown query (ISO 8601 format) workforce_ids: type: array items: type: string description: Optional array of workforce display IDs to filter breakdown by agent_ids: type: array items: type: string description: Optional array of agent IDs to filter breakdown by required: - start_date - end_date additionalProperties: false GetAgentUsageTimeseriesInput: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start date (YYYY-MM-DD) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end date (YYYY-MM-DD) agent_ids: type: array items: type: string minLength: 1 maxLength: 256 maxItems: 1000 description: Optional list of agent display IDs to filter by. Duplicates are ignored. Maximum 1000 IDs. required: - start_date - end_date additionalProperties: false GetToolUsageBreakdownInput: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start date (YYYY-MM-DD) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end date (YYYY-MM-DD) tool_ids: type: array items: type: string minLength: 1 maxLength: 256 maxItems: 1000 description: Optional list of tool display IDs to filter by. Duplicates are ignored. Maximum 1000 IDs. required: - start_date - end_date additionalProperties: false GetWorkforceUsageTimeseriesOutput: type: object properties: meta: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start of the data range covered by this response (as specified by the request) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end of the data range covered by this response (as specified by the request) interval: type: string enum: - day - total description: Bucket granularity. `day` on time-bucketed endpoints (timeseries / breakdown_timeseries); `total` on plain breakdown endpoints (single bucket covering the whole window). Always present. timezone: type: string description: IANA timezone the daily buckets were materialized in (mirrors the org's internally set timezone). entity: type: string enum: - tool - agent - workforce - project - model description: Entity scope of the response. Each v2 analytics endpoint owns a single entity type; `project` denotes project-wide rollups; `model` denotes AI model usage rollups. required: - start_date - end_date - interval - timezone - entity additionalProperties: false rows: type: array items: type: object properties: date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Bucket date tasks: type: number description: Count of tasks across all selected workforces actions: type: number description: Count of actions across all selected workforces agent_model_credits: type: number description: Credits consumed by agents' own model calls inside the selected workforces for the bucket. Pairs with tool_credits (disjoint; mirrors grand-totals semantics so workforce-row credits aren't double-counted). tool_credits: type: number description: Credits consumed by tools agents ran inside the selected workforces for the bucket. Pairs with agent_model_credits. error_rate: type: number description: Fraction of tasks not completed (success = state_completed), in [0, 1]; 0 when there are no tasks in the bucket required: - date - tasks - actions - agent_model_credits - tool_credits - error_rate additionalProperties: false totals: type: object properties: grand: type: object properties: tasks: type: number description: Count of tasks (the workforce orchestration unit; workforce primary metric) actions: type: number description: Count of actions agent_model_credits: type: number description: Credits consumed by agents' own model calls inside the selected workforces. Pairs with tool_credits (disjoint; avoids workforce-row double-count). tool_credits: type: number description: Credits consumed by tools agents ran inside the selected workforces. Pairs with agent_model_credits. credits_per_task: type: number description: (agent_model_credits + tool_credits) / tasks ratio. 0 when tasks is 0. error_rate: type: number description: Fraction of tasks not completed (success = state_completed), in [0, 1]; 0 when tasks is 0 required: - tasks - actions - agent_model_credits - tool_credits - credits_per_task - error_rate additionalProperties: false required: - grand additionalProperties: false required: - meta - rows - totals additionalProperties: false GetAgentUsageBreakdownOutput: type: object properties: meta: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start of the data range covered by this response (as specified by the request) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end of the data range covered by this response (as specified by the request) interval: type: string enum: - day - total description: Bucket granularity. `day` on time-bucketed endpoints (timeseries / breakdown_timeseries); `total` on plain breakdown endpoints (single bucket covering the whole window). Always present. timezone: type: string description: IANA timezone the daily buckets were materialized in (mirrors the org's internally set timezone). entity: type: string enum: - tool - agent - workforce - project - model description: Entity scope of the response. Each v2 analytics endpoint owns a single entity type; `project` denotes project-wide rollups; `model` denotes AI model usage rollups. required: - start_date - end_date - interval - timezone - entity additionalProperties: false rows: type: array items: type: object properties: entity_id: type: string description: Agent display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' tasks: type: number description: Count of tasks (the agent orchestration unit; agent primary metric) error_rate: type: number description: Fraction of tasks not in a success state (success = state_completed + state_idle), in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_task: type: number description: credits / tasks ratio. 0 when tasks is 0. children: type: array items: type: object properties: entity_id: type: string description: Tool display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' runs: type: number description: Count of runs error_rate: type: number description: Fraction of tool runs that did not complete, in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_run: type: number description: credits / runs ratio. 0 when runs is 0. required: - entity_id - name - description - icon - runs - error_rate - actions - credits - credits_per_run additionalProperties: false description: Present only when the request `children=yes`. Per-tool rollups stitched at the agent grain — one entry per tool the agent invoked over the window. Tool-grained metrics are runs-keyed (primary metric per entity). required: - entity_id - name - description - icon - tasks - error_rate - actions - credits - credits_per_task additionalProperties: false totals: type: object properties: grand: type: object properties: tasks: type: number description: Count of tasks (the agent orchestration unit; agent primary metric) actions: type: number description: Count of actions credits: type: number description: Total credits consumed agent_model_credits: type: number description: Credits consumed by the agent's own model calls. Subset of `credits`; the remainder is from tools the agent ran. credits_per_task: type: number description: credits / tasks ratio. 0 when tasks is 0. error_rate: type: number description: Fraction of tasks not in a success state (success = state_completed + state_idle), in [0, 1]; 0 when tasks is 0 required: - tasks - actions - credits - agent_model_credits - credits_per_task - error_rate additionalProperties: false required: - grand additionalProperties: false required: - meta - rows - totals additionalProperties: false GetAgentErrorsInput: type: object properties: start_date: description: Start date for the errors query (ISO 8601 format) end_date: description: End date for the errors query (ISO 8601 format) agent_ids: type: array items: type: string description: Optional array of agent IDs to filter errors by. If omitted, returns errors for all agents in the project. page_size: type: number minimum: 1 maximum: 100 default: 20 description: 'Number of errors per page (default: 20, max: 100)' cursor: type: object properties: timestamp: type: string description: ISO 8601 timestamp of the last error returned required: - timestamp additionalProperties: false required: - start_date - end_date additionalProperties: false GetModelUsageBreakdownInput: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start date (YYYY-MM-DD) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end date (YYYY-MM-DD) model_ids: type: array items: type: string minLength: 1 maxLength: 256 maxItems: 1000 description: Optional list of model identifiers (e.g. 'gpt-4o', 'claude-3-7-sonnet-latest') to filter by. Duplicates are ignored. Maximum 1000. contexts: type: array items: type: string enum: - tool - agent_model - agent_tool - workforce_tool - workforce_agent_tool - workforce_agent_model description: Where the AI model call happened in the workforce/agent/tool tree. `tool` = standalone tool run, no agent or workforce; `agent_model` = an agent's own AI model call, no tool or workforce; `agent_tool` = a tool inside an agent task, no workforce; `workforce_tool` = a workforce-driven tool with no agent; `workforce_agent_tool` = a tool inside an agent task inside a workforce; `workforce_agent_model` = an agent's own AI model call inside a workforce, no tool. maxItems: 6 description: Optional list of context buckets to include. Duplicates are ignored. Defaults to all six contexts when omitted. required: - start_date - end_date additionalProperties: false GetActionBreakdownOutput: type: object properties: actions: type: array items: type: object properties: action_id: type: string description: Action/Tool display ID action_name: type: string description: Action/Tool name execution_count: type: number description: Total number of executions for this action error_rate: type: number description: Error rate percentage (0-100) error_count: type: number description: Number of errored executions (the raw count behind error_rate). credits_used: type: number description: Total credits used by this action top_agents: type: array items: type: object properties: agent_id: type: string description: Agent display ID agent_name: type: string description: Agent name (fallback to default if null) task_count: type: number description: Number of tasks this agent executed using this action required: - agent_id - agent_name - task_count additionalProperties: false description: Top 3 agents that use this action most maxItems: 3 required: - action_id - action_name - execution_count - error_rate - credits_used - top_agents additionalProperties: false description: Action breakdown data required: - actions additionalProperties: false GetAgentModelBreakdownInput: type: object properties: start_date: description: Start date for the model breakdown query (ISO 8601 format) end_date: description: End date for the model breakdown query (ISO 8601 format) agent_ids: type: array items: type: string description: Optional array of agent IDs to filter usage by required: - start_date - end_date additionalProperties: false GetErrorTimeSeriesOutput: type: object properties: errors: type: array items: type: object properties: source_type: type: string enum: - tool - agent source_id: type: string source_name: type: - string - 'null' source_emoji: type: - string - 'null' source_context: type: object properties: agent_id: type: - string - 'null' agent_name: type: - string - 'null' conversation_id: type: - string - 'null' workforce_id: type: - string - 'null' workforce_name: type: - string - 'null' workforce_task_id: type: - string - 'null' additionalProperties: false error_time: {} error_message: type: string required: - source_type - source_id - source_context - error_time - error_message additionalProperties: false has_more: type: boolean next_cursor: type: - string - 'null' time_series: type: array items: type: object properties: date: type: string count: type: number required: - date - count required: - errors - has_more additionalProperties: false GetToolUsageTimeseriesInput: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start date (YYYY-MM-DD) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end date (YYYY-MM-DD) tool_ids: type: array items: type: string minLength: 1 maxLength: 256 maxItems: 1000 description: Optional list of tool display IDs to filter by. Duplicates are ignored. Maximum 1000 IDs. required: - start_date - end_date additionalProperties: false GetToolUsageTimeseriesOutput: type: object properties: meta: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start of the data range covered by this response (as specified by the request) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end of the data range covered by this response (as specified by the request) interval: type: string enum: - day - total description: Bucket granularity. `day` on time-bucketed endpoints (timeseries / breakdown_timeseries); `total` on plain breakdown endpoints (single bucket covering the whole window). Always present. timezone: type: string description: IANA timezone the daily buckets were materialized in (mirrors the org's internally set timezone). entity: type: string enum: - tool - agent - workforce - project - model description: Entity scope of the response. Each v2 analytics endpoint owns a single entity type; `project` denotes project-wide rollups; `model` denotes AI model usage rollups. required: - start_date - end_date - interval - timezone - entity additionalProperties: false rows: type: array items: type: object properties: date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Bucket date runs: type: number description: Count of runs across all selected tools actions: type: number description: Count of actions across all selected tools credits: type: number description: Total credits consumed for the bucket error_rate: type: number description: Fraction of runs that did not complete, in [0, 1]; 0 when there are no runs in the bucket required: - date - runs - actions - credits - error_rate additionalProperties: false totals: type: object properties: grand: type: object properties: runs: type: number description: Count of runs actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_run: type: number description: credits / runs ratio. 0 when runs is 0. error_rate: type: number description: Fraction of runs that did not complete, in [0, 1]; 0 when runs is 0 required: - runs - actions - credits - credits_per_run - error_rate additionalProperties: false required: - grand additionalProperties: false required: - meta - rows - totals additionalProperties: false GetActionBreakdownInput: type: object properties: start_date: description: Start date for the action breakdown query (ISO 8601 format) end_date: description: End date for the action breakdown query (ISO 8601 format) action_ids: type: array items: type: string description: Optional array of action IDs to filter breakdown by agent_ids: type: array items: type: string description: Optional array of agent IDs to filter breakdown by required: - start_date - end_date additionalProperties: false GetToolUsageBreakdownTimeseriesInput: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start date (YYYY-MM-DD) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end date (YYYY-MM-DD) tool_ids: type: array items: type: string minLength: 1 maxLength: 256 maxItems: 1000 description: Optional list of tool display IDs to filter by. Duplicates are ignored. Maximum 1000 IDs. required: - start_date - end_date additionalProperties: false GetToolUsageBreakdownTimeseriesOutput: type: object properties: meta: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start of the data range covered by this response (as specified by the request) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end of the data range covered by this response (as specified by the request) interval: type: string enum: - day - total description: Bucket granularity. `day` on time-bucketed endpoints (timeseries / breakdown_timeseries); `total` on plain breakdown endpoints (single bucket covering the whole window). Always present. timezone: type: string description: IANA timezone the daily buckets were materialized in (mirrors the org's internally set timezone). entity: type: string enum: - tool - agent - workforce - project - model description: Entity scope of the response. Each v2 analytics endpoint owns a single entity type; `project` denotes project-wide rollups; `model` denotes AI model usage rollups. required: - start_date - end_date - interval - timezone - entity additionalProperties: false entities: type: array items: type: object properties: entity_id: type: string description: Tool display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' runs: type: number description: Count of runs error_rate: type: number description: Fraction of runs that did not complete, in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_run: type: number description: credits / runs ratio. 0 when runs is 0. required: - entity_id - name - description - icon - runs - error_rate - actions - credits - credits_per_run additionalProperties: false series: type: array items: type: object properties: entity_id: type: string description: Tool display ID; joins to `entities[].entity_id` date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Bucket date runs: type: number description: Count of runs credits: type: number description: Total credits for the day actions: type: number description: Count of actions error_rate: type: number description: Fraction of runs that did not complete, in [0, 1]; 0 when there are no runs in the bucket required: - entity_id - date - runs - credits - actions - error_rate additionalProperties: false totals: type: object properties: grand: type: object properties: runs: type: number description: Count of runs actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_run: type: number description: credits / runs ratio. 0 when runs is 0. error_rate: type: number description: Fraction of runs that did not complete, in [0, 1]; 0 when runs is 0 required: - runs - actions - credits - credits_per_run - error_rate additionalProperties: false by_date: type: array items: type: object properties: date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Bucket date runs: type: number description: Count of runs across all selected tools actions: type: number description: Count of actions across all selected tools credits: type: number description: Total credits consumed for the bucket error_rate: type: number description: Fraction of runs that did not complete, in [0, 1]; 0 when there are no runs in the bucket required: - date - runs - actions - credits - error_rate additionalProperties: false required: - grand - by_date additionalProperties: false required: - meta - entities - series - totals additionalProperties: false GetAgentModelBreakdownOutput: type: object properties: usage: type: array items: type: object properties: model_name: type: string description: LLM model display name (falls back to model ID if display name not available) credits: type: number description: Credits used for this model input_tokens: type: number description: Total input tokens for this model output_tokens: type: number description: Total output tokens for this model vendor_icon_url: type: string description: URL for the model vendor's icon/logo required: - model_name - credits - input_tokens - output_tokens additionalProperties: false description: Model breakdown data required: - usage additionalProperties: false GetWorkforceUsageBreakdownOutput: type: object properties: meta: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start of the data range covered by this response (as specified by the request) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end of the data range covered by this response (as specified by the request) interval: type: string enum: - day - total description: Bucket granularity. `day` on time-bucketed endpoints (timeseries / breakdown_timeseries); `total` on plain breakdown endpoints (single bucket covering the whole window). Always present. timezone: type: string description: IANA timezone the daily buckets were materialized in (mirrors the org's internally set timezone). entity: type: string enum: - tool - agent - workforce - project - model description: Entity scope of the response. Each v2 analytics endpoint owns a single entity type; `project` denotes project-wide rollups; `model` denotes AI model usage rollups. required: - start_date - end_date - interval - timezone - entity additionalProperties: false rows: type: array items: type: object properties: entity_id: type: string description: Workforce display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' tasks: type: number description: Count of tasks (the workforce orchestration unit; workforce primary metric) error_rate: type: number description: Fraction of tasks not completed (workforce success = state_completed only; state_idle does not apply to workforces), in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_task: type: number description: credits / tasks ratio. 0 when tasks is 0. children: type: array items: type: object properties: entity_id: type: string description: Agent display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' tasks: type: number description: Count of tasks (the agent orchestration unit; agent primary metric) error_rate: type: number description: Fraction of tasks not in a success state (agent success = state_completed + state_idle), in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_task: type: number description: credits / tasks ratio. 0 when tasks is 0. children: type: array items: type: object properties: entity_id: type: string description: Tool display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' runs: type: number description: Count of runs error_rate: type: number description: Fraction of tool runs that did not complete, in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_run: type: number description: credits / runs ratio. 0 when runs is 0. required: - entity_id - name - description - icon - runs - error_rate - actions - credits - credits_per_run additionalProperties: false description: Present only when the request `children=yes`. Per-tool rollups stitched at the agent grain — one entry per tool the agent invoked within this workforce over the window. Tool-grained children remain runs-keyed (primary metric per entity). required: - entity_id - name - description - icon - tasks - error_rate - actions - credits - credits_per_task additionalProperties: false description: Present only when the request `children=yes`. Per-agent rollups stitched at the workforce grain. Each agent child may carry its own nested `children` array of per-tool rollups (the full workforce -> agent -> tool tree). required: - entity_id - name - description - icon - tasks - error_rate - actions - credits - credits_per_task additionalProperties: false totals: type: object properties: grand: type: object properties: tasks: type: number description: Count of tasks (the workforce orchestration unit; workforce primary metric) actions: type: number description: Count of actions agent_model_credits: type: number description: Credits consumed by agents' own model calls inside the selected workforces. Pairs with tool_credits (disjoint; mirrors project-level semantics so workforce/agent/tool credits aren't double-counted). tool_credits: type: number description: Credits consumed by tools run inside the selected workforces. Pairs with agent_model_credits. credits_per_task: type: number description: (agent_model_credits + tool_credits) / tasks ratio. 0 when tasks is 0. error_rate: type: number description: Fraction of tasks not completed (workforce success = state_completed only; no state_idle exists at the workforce level), in [0, 1]; 0 when tasks is 0 required: - tasks - actions - agent_model_credits - tool_credits - credits_per_task - error_rate additionalProperties: false required: - grand additionalProperties: false required: - meta - rows - totals additionalProperties: false GetMetricsInput: type: object properties: metric: type: string enum: - total_tasks - tasks_to_review - total_credits - total_actions - daily_active_users - tasks_escalated description: Type of metric to retrieve group_by: type: string enum: - day description: Time grouping for the data points default: day start_date: description: Start date for the metrics query (ISO 8601 format) end_date: description: End date for the metrics query (ISO 8601 format) agent_ids: type: array items: type: string description: Optional array of agent IDs to filter metrics by required: - metric - start_date - end_date additionalProperties: false GetAgentUsageBreakdownTimeseriesInput: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start date (YYYY-MM-DD) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end date (YYYY-MM-DD) agent_ids: type: array items: type: string minLength: 1 maxLength: 256 maxItems: 1000 description: Optional list of agent display IDs to filter by. Duplicates are ignored. Maximum 1000 IDs. children: type: string enum: - 'yes' - 'no' description: When `yes`, breakdown-shaped endpoints stitch descendant rollups onto each parent entity's `children` array. Default `no`. required: - start_date - end_date additionalProperties: false GetWorkforceUsageBreakdownInput: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start date (YYYY-MM-DD) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end date (YYYY-MM-DD) workforce_ids: type: array items: type: string minLength: 1 maxLength: 256 maxItems: 1000 description: Optional list of workforce display IDs to filter by. Duplicates are ignored. Maximum 1000 IDs. children: type: string enum: - 'yes' - 'no' description: When `yes`, breakdown-shaped endpoints stitch descendant rollups onto each parent entity's `children` array. Default `no`. required: - start_date - end_date additionalProperties: false GetWorkforceUsageBreakdownTimeseriesInput: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start date (YYYY-MM-DD) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end date (YYYY-MM-DD) workforce_ids: type: array items: type: string minLength: 1 maxLength: 256 maxItems: 1000 description: Optional list of workforce display IDs to filter by. Duplicates are ignored. Maximum 1000 IDs. children: type: string enum: - 'yes' - 'no' description: When `yes`, breakdown-shaped endpoints stitch descendant rollups onto each parent entity's `children` array. Default `no`. required: - start_date - end_date additionalProperties: false GetConcurrencyTimeSeriesByOrgIdOutput: type: object properties: results: type: array items: type: object properties: timestamp: type: string description: ISO 8601 timestamp for the bucket value: type: number description: Aggregated concurrency value queued: type: number description: Jobs waiting for a slot at the bucket's peak instant. Present only when queue-depth reporting is enabled. concurrency_limit: type: - number - 'null' description: Org concurrency limit in effect at the bucket's peak instant; null for buckets predating queue-depth logging. Present only when queue-depth reporting is enabled. required: - timestamp - value additionalProperties: false by_project: type: array description: Per-project series, present only when group_by="project" is set. `usage` is the concurrency slots held by the project at the org's peak instant in the bucket. `queued` is reported as 0 until queue-depth logging lands, at which point it is populated from real data without a contract change. When the caller is not an org admin, projects other than their own are collapsed into a single anonymised bucket with `is_redacted=true` and null `project_id`/`project_name`. items: type: object properties: project_id: type: - string - 'null' description: Project display ID, or null for the anonymised bucket that aggregates projects the caller may not see. project_name: type: - string - 'null' is_redacted: type: boolean description: True for the anonymised bucket aggregating projects the caller is not permitted to see individually. results: type: array items: type: object properties: timestamp: type: string usage: type: number queued: type: number required: - timestamp - usage - queued additionalProperties: false required: - project_id - project_name - is_redacted - results additionalProperties: false metadata: type: object properties: granularity: type: string enum: - 15m - 4h - 1d aggregation: type: string enum: - max - avg start_date: type: string end_date: type: string concurrency_limit: type: number required: - granularity - aggregation - start_date - end_date - concurrency_limit additionalProperties: false required: - results - metadata additionalProperties: false GetToolUsageBreakdownOutput: type: object properties: meta: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start of the data range covered by this response (as specified by the request) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end of the data range covered by this response (as specified by the request) interval: type: string enum: - day - total description: Bucket granularity. `day` on time-bucketed endpoints (timeseries / breakdown_timeseries); `total` on plain breakdown endpoints (single bucket covering the whole window). Always present. timezone: type: string description: IANA timezone the daily buckets were materialized in (mirrors the org's internally set timezone). entity: type: string enum: - tool - agent - workforce - project - model description: Entity scope of the response. Each v2 analytics endpoint owns a single entity type; `project` denotes project-wide rollups; `model` denotes AI model usage rollups. required: - start_date - end_date - interval - timezone - entity additionalProperties: false rows: type: array items: type: object properties: entity_id: type: string description: Tool display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' runs: type: number description: Count of runs error_rate: type: number description: Fraction of runs that did not complete, in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_run: type: number description: credits / runs ratio. 0 when runs is 0. required: - entity_id - name - description - icon - runs - error_rate - actions - credits - credits_per_run additionalProperties: false totals: type: object properties: grand: type: object properties: runs: type: number description: Count of runs actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_run: type: number description: credits / runs ratio. 0 when runs is 0. error_rate: type: number description: Fraction of runs that did not complete, in [0, 1]; 0 when runs is 0 required: - runs - actions - credits - credits_per_run - error_rate additionalProperties: false required: - grand additionalProperties: false required: - meta - rows - totals additionalProperties: false GetProjectUsageTimeseriesInput: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start date (YYYY-MM-DD) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end date (YYYY-MM-DD) required: - start_date - end_date additionalProperties: false GetTaskTimelineOutput: type: object properties: time_series: type: array items: type: object properties: date: type: string count: type: number required: - date - count additionalProperties: false required: - time_series additionalProperties: false GetAgentUsageBreakdownInput: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start date (YYYY-MM-DD) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end date (YYYY-MM-DD) agent_ids: type: array items: type: string minLength: 1 maxLength: 256 maxItems: 1000 description: Optional list of agent display IDs to filter by. Duplicates are ignored. Maximum 1000 IDs. children: type: string enum: - 'yes' - 'no' description: When `yes`, breakdown-shaped endpoints stitch descendant rollups onto each parent entity's `children` array. Default `no`. required: - start_date - end_date additionalProperties: false GetAgentBreakdownInput: type: object properties: start_date: description: Start date for the agent breakdown query (ISO 8601 format) end_date: description: End date for the agent breakdown query (ISO 8601 format) agent_ids: type: array items: type: string description: Optional array of agent IDs to filter breakdown by required: - start_date - end_date additionalProperties: false GetProjectUsageTimeseriesOutput: type: object properties: meta: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start of the data range covered by this response (as specified by the request) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end of the data range covered by this response (as specified by the request) interval: type: string enum: - day - total description: Bucket granularity. `day` on time-bucketed endpoints (timeseries / breakdown_timeseries); `total` on plain breakdown endpoints (single bucket covering the whole window). Always present. timezone: type: string description: IANA timezone the daily buckets were materialized in (mirrors the org's internally set timezone). entity: type: string enum: - tool - agent - workforce - project - model description: Entity scope of the response. Each v2 analytics endpoint owns a single entity type; `project` denotes project-wide rollups; `model` denotes AI model usage rollups. required: - start_date - end_date - interval - timezone - entity additionalProperties: false rows: type: array items: type: object properties: date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Bucket date agent_tasks: type: number workforce_tasks: type: number tool_runs: type: number actions: type: number tool_credits: type: number description: Per-day credits consumed by tool execution, including AI model calls made inside tools. agent_model_credits: type: number description: Per-day credits consumed by agents' own model calls. required: - date - agent_tasks - workforce_tasks - tool_runs - actions - tool_credits - agent_model_credits additionalProperties: false totals: type: object properties: grand: type: object properties: agent_tasks: type: number workforce_tasks: type: number tool_runs: type: number actions: type: number tool_credits: type: number description: Credits consumed by tool execution, including AI model calls made inside tools. agent_model_credits: type: number description: Credits consumed by agents' own AI model calls (the agent's reasoning model). Tool costs the agent triggered are reported under `tool_credits`. required: - agent_tasks - workforce_tasks - tool_runs - actions - tool_credits - agent_model_credits additionalProperties: false required: - grand additionalProperties: false required: - meta - rows - totals additionalProperties: false GetMetricsOutput: type: object properties: metric_type: type: string enum: - total_tasks - tasks_to_review - total_credits - total_actions - daily_active_users - tasks_escalated description: The metric type this data represents data_points: type: array items: type: object properties: timestamp: description: Timestamp for this data point value: type: number description: Metric value at this timestamp grouped_by: type: string enum: - day description: Time grouping for the data points required: - timestamp - value - grouped_by additionalProperties: false description: Time-series data points for the metric summary: type: object properties: total: type: number description: Sum of all values in the period average: type: number description: Average value in the period min_value: type: number description: Minimum value in the period max_value: type: number description: Maximum value in the period required: - total - average - min_value - max_value additionalProperties: false description: Summary statistics for the metric required: - metric_type - data_points - summary additionalProperties: false GetLiveJobsByOrgIdOutput: type: object properties: results: type: array items: type: object properties: job_id: type: - string - 'null' state: type: string enum: - queued - running kind: type: string enum: - agent - workforce - tool - unknown status: type: string enum: - pending - pending_dispatch - dispatched - running title: type: - string - 'null' task_title: type: - string - 'null' executor_type: type: - string - 'null' source: type: object properties: name: type: - string - 'null' emoji: type: - string - 'null' trigger_source: type: - string - 'null' description: Source trigger type when available from the backing conversation, for example webhook or email. Null for manual/API runs, tool-only jobs, unknown jobs, or redacted jobs. required: - name - emoji - trigger_source additionalProperties: false agent_display_id: type: - string - 'null' workforce_display_id: type: - string - 'null' workforce_task_display_id: type: - string - 'null' conversation_display_id: type: - string - 'null' project_display_id: type: - string - 'null' project_name: type: - string - 'null' created_at: type: string format: date-time age_seconds: type: number description: Seconds since the scheduler job was created. queue_rank: type: - number - 'null' priority_tier: type: string enum: - high - medium - low - untiered description: Resolved priority tier stamped at enqueue; 'untiered' when no tier was resolved. is_redacted: type: boolean description: True when project-scoped callers can see org pressure but not this row's task details. required: - job_id - state - kind - status - title - task_title - executor_type - source - agent_display_id - workforce_display_id - workforce_task_display_id - conversation_display_id - project_display_id - project_name - created_at - age_seconds - queue_rank - priority_tier - is_redacted additionalProperties: false metadata: type: object properties: active_count: type: number running_count: type: number queued_count: type: number next_cursor: type: - string - 'null' truncated: type: boolean concurrency_limit: type: number required: - active_count - running_count - queued_count - next_cursor - truncated - concurrency_limit additionalProperties: false required: - results - metadata additionalProperties: false GetAgentUsageBreakdownTimeseriesOutput: type: object properties: meta: type: object properties: start_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive start of the data range covered by this response (as specified by the request) end_date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Inclusive end of the data range covered by this response (as specified by the request) interval: type: string enum: - day - total description: Bucket granularity. `day` on time-bucketed endpoints (timeseries / breakdown_timeseries); `total` on plain breakdown endpoints (single bucket covering the whole window). Always present. timezone: type: string description: IANA timezone the daily buckets were materialized in (mirrors the org's internally set timezone). entity: type: string enum: - tool - agent - workforce - project - model description: Entity scope of the response. Each v2 analytics endpoint owns a single entity type; `project` denotes project-wide rollups; `model` denotes AI model usage rollups. required: - start_date - end_date - interval - timezone - entity additionalProperties: false entities: type: array items: type: object properties: entity_id: type: string description: Agent display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' tasks: type: number description: Count of tasks (the agent orchestration unit; agent primary metric) error_rate: type: number description: Fraction of tasks not in a success state (success = state_completed + state_idle), in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_task: type: number description: credits / tasks ratio. 0 when tasks is 0. children: type: array items: type: object properties: entity_id: type: string description: Tool display ID (UUID) name: type: - string - 'null' description: type: - string - 'null' icon: type: - string - 'null' runs: type: number description: Count of runs error_rate: type: number description: Fraction of tool runs that did not complete, in [0, 1] actions: type: number description: Count of actions credits: type: number description: Total credits consumed credits_per_run: type: number description: credits / runs ratio. 0 when runs is 0. required: - entity_id - name - description - icon - runs - error_rate - actions - credits - credits_per_run additionalProperties: false description: Present only when the request `children=yes`. Per-tool rollups stitched at the agent grain — one entry per tool the agent invoked over the window. Tool-grained metrics are runs-keyed (primary metric per entity). required: - entity_id - name - description - icon - tasks - error_rate - actions - credits - credits_per_task additionalProperties: false series: type: array items: type: object properties: entity_id: type: string description: Agent display ID; joins to `entities[].entity_id` date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Bucket date tasks: type: number description: Count of tasks for this (agent, day) credits: type: number description: Total credits for the day actions: type: number description: Count of actions error_rate: type: number description: Fraction of tasks not in a success state, in [0, 1]; 0 when there are no tasks required: - entity_id - date - tasks - credits - actions - error_rate additionalProperties: false totals: type: object properties: grand: type: object properties: tasks: type: number description: Count of tasks (the agent orchestration unit; agent primary metric) actions: type: number description: Count of actions credits: type: number description: Total credits consumed agent_model_credits: type: number description: Credits consumed by the agent's own model calls. Subset of `credits`; the remainder is from tools the agent ran. credits_per_task: type: number description: credits / tasks ratio. 0 when tasks is 0. error_rate: type: number description: Fraction of tasks not in a success state (success = state_completed + state_idle), in [0, 1]; 0 when tasks is 0 required: - tasks - actions - credits - agent_model_credits - credits_per_task - error_rate additionalProperties: false by_date: type: array items: type: object properties: date: type: string format: date pattern: ^\d{4}-\d{2}-\d{2}$ description: Bucket date tasks: type: number description: Count of tasks across all selected agents actions: type: number description: Count of actions across all selected agents credits: type: number description: Total credits consumed for the bucket error_rate: type: number description: Fraction of tasks not in a success state, in [0, 1]; 0 when there are no tasks in the bucket required: - date - tasks - actions - credits - error_rate additionalProperties: false required: - grand - by_date additionalProperties: false required: - meta - entities - series - totals additionalProperties: false securitySchemes: AuthorizationHeader: type: apiKey in: header name: Authorization description: 'Authorization credentials. Header authorization should be in the form of: project:api_key'