# Generated by API Evangelist (build-phrasing.py). Our phrasing, not observed demand. overlay: 1.0.0 info: title: API Evangelist conversational phrasing for Parallel Tasks API version: 1.0.0 extends: openapi/parallel-tasks-api-openapi.yml actions: - target: $.info update: x-apievangelist-phrasing: method: generated generated: '2026-10-01' generator: build-phrasing.py label: Generated by API Evangelist operations: 12 - target: $.paths['/v1/tasks/groups'].post update: x-apievangelist-phrasing: intent: Create a task group to batch runs effect: write questions: - How do I group many task runs together so I can track them as one batch? - Can I attach my own metadata to a task group when I create it? instructions: - text: Create a new task group. - text: Create a task group tagged with metadata {metadata}. slots: metadata: requestBody.metadata method: generated generated: '2026-10-01' - target: $.paths['/v1/tasks/groups/{taskgroup_id}'].get update: x-apievangelist-phrasing: intent: Get aggregated status of a task group effect: read questions: - How many runs in my task group have finished versus are still running? - Can I see the overall progress of a batch of task runs at a glance? instructions: - text: Show the aggregated status of task group {taskgroup_id}. slots: taskgroup_id: path.taskgroup_id - text: Summarize progress across all runs in task group {taskgroup_id}. slots: taskgroup_id: path.taskgroup_id method: generated generated: '2026-10-01' - target: $.paths['/v1/tasks/groups/{taskgroup_id}/events'].get update: x-apievangelist-phrasing: intent: Stream status events for a task group effect: read questions: - Can I get live notifications as runs in a task group complete? - How long does a task group event stream stay open? instructions: - text: Stream status updates and run completions for task group {taskgroup_id}. slots: taskgroup_id: path.taskgroup_id - text: Resume task group {taskgroup_id} events after event {last_event_id}. slots: taskgroup_id: path.taskgroup_id last_event_id: query.last_event_id method: generated generated: '2026-10-01' - target: $.paths['/v1/tasks/groups/{taskgroup_id}/runs'].get update: x-apievangelist-phrasing: intent: Fetch the runs in a task group effect: read questions: - How do I get every run in a task group along with its inputs and outputs? - Can I fetch only the task group runs with a particular status? - Is the stream of task group runs resumable if it gets interrupted? instructions: - text: Fetch all runs in task group {taskgroup_id} with their outputs. slots: taskgroup_id: path.taskgroup_id - text: List the {status} runs in task group {taskgroup_id}. slots: status: query.status taskgroup_id: path.taskgroup_id method: generated generated: '2026-10-01' - target: $.paths['/v1/tasks/groups/{taskgroup_id}/runs'].post update: x-apievangelist-phrasing: intent: Add a batch of task runs to a group effect: write questions: - How many task runs can I add to a task group in a single request? - Can I set a default task spec for all the runs I add to a group? instructions: - text: 'Add these runs to task group {taskgroup_id}: {inputs}.' slots: taskgroup_id: path.taskgroup_id inputs: requestBody.inputs - text: Queue {inputs} in task group {taskgroup_id} using default spec {default_task_spec}. slots: inputs: requestBody.inputs taskgroup_id: path.taskgroup_id default_task_spec: requestBody.default_task_spec method: generated generated: '2026-10-01' - target: $.paths['/v1/tasks/groups/{taskgroup_id}/runs/{run_id}'].get update: x-apievangelist-phrasing: intent: Check one run inside a task group effect: read questions: - Can I look up the status of a single run within a specific task group? - Where do I check one run's status when I only know its group and run IDs? instructions: - text: Check the status of run {run_id} in task group {taskgroup_id}. slots: run_id: path.run_id taskgroup_id: path.taskgroup_id - text: Look up group {taskgroup_id} member run {run_id}. slots: taskgroup_id: path.taskgroup_id run_id: path.run_id method: generated generated: '2026-10-01' - target: $.paths['/v1/tasks/runs'].post update: x-apievangelist-phrasing: intent: Start a web research task run effect: write questions: - How do I start a single research task with Parallel and get structured output? - Which processor should a task run use, and can I restrict which domains it searches? - Can a task run call my own MCP servers or notify a webhook when it's done? instructions: - text: Start a task run on the {processor} processor with input {input}. slots: processor: requestBody.processor input: requestBody.input - text: Run {input} with processor {processor} using task spec {task_spec}. slots: input: requestBody.input processor: requestBody.processor task_spec: requestBody.task_spec - text: Launch a {processor} task for {input} and call {webhook} when it completes. slots: processor: requestBody.processor input: requestBody.input webhook: requestBody.webhook method: generated generated: '2026-10-01' - target: $.paths['/v1/tasks/runs/{run_id}'].get update: x-apievangelist-phrasing: intent: Check a task run's status effect: read questions: - Is my task run still queued or has it finished? - How can I poll a task run's status without waiting for the result? instructions: - text: Check the status of task run {run_id}. slots: run_id: path.run_id - text: Tell me if task run {run_id} is done yet. slots: run_id: path.run_id method: generated generated: '2026-10-01' - target: $.paths['/v1/tasks/runs/{run_id}/events'].get update: x-apievangelist-phrasing: intent: Stream progress events for a task run effect: read questions: - Can I watch a task run's progress updates as they happen on the current v1 endpoint? - Will I get fewer progress events if I didn't enable events when creating the run? instructions: - text: Stream progress events for task run {run_id} using the v1 endpoint. slots: run_id: path.run_id - text: Follow task run {run_id} state changes live via v1. slots: run_id: path.run_id method: generated generated: '2026-10-01' - target: $.paths['/v1/tasks/runs/{run_id}/input'].get update: x-apievangelist-phrasing: intent: Get the input a task run was given effect: read questions: - What input did I submit for a particular task run? - Can I retrieve the original prompt or JSON a task run started with? instructions: - text: Show the input of task run {run_id}. slots: run_id: path.run_id - text: Get what task run {run_id} was asked to research. slots: run_id: path.run_id method: generated generated: '2026-10-01' - target: $.paths['/v1/tasks/runs/{run_id}/result'].get update: x-apievangelist-phrasing: intent: Get the output of a completed task run effect: read questions: - How do I get the final output of a task run once it completes? - Does fetching a task run's result block until it's finished, and can I set a timeout? instructions: - text: Get the result of task run {run_id}. slots: run_id: path.run_id - text: Wait up to {timeout} seconds for task run {run_id} output. slots: timeout: query.timeout run_id: path.run_id method: generated generated: '2026-10-01' - target: $.paths['/v1beta/tasks/runs/{run_id}/events'].get update: x-apievangelist-phrasing: intent: Stream task run events (legacy v1beta) effect: read questions: - Is there still a v1beta endpoint for streaming task run events? - Can older beta integrations keep streaming task run progress from the beta path? instructions: - text: Stream events for task run {run_id} from the v1beta endpoint. slots: run_id: path.run_id - text: Use the legacy beta stream to follow task run {run_id}. slots: run_id: path.run_id method: generated generated: '2026-10-01'