# Generated by API Evangelist (build-phrasing.py). Our phrasing, not observed demand.
overlay: 1.0.0
info:
  title: API Evangelist conversational phrasing for Browser Use Public API v2 Tasks API
  version: 1.0.0
extends: openapi/browser-use-tasks-api-openapi.yml
actions:
- target: $.info
  update:
    x-apievangelist-phrasing:
      method: generated
      generated: '2026-09-26'
      generator: build-phrasing.py
      label: Generated by API Evangelist
      operations: 6
- target: $.paths['/tasks'].get
  update:
    x-apievangelist-phrasing:
      intent: List AI agent tasks
      effect: read
      questions:
      - Which browser agent tasks have I run, and which are still in progress?
      - Can I list only the tasks from one session or within a time window?
      instructions:
      - text: List the tasks in session {session_id} with status {status}.
        slots:
          session_id: query.sessionId
          status: query.filterBy
      - text: Show my agent tasks created after {after} and before {before}.
        slots:
          after: query.after
          before: query.before
      method: generated
      generated: '2026-09-26'
- target: $.paths['/tasks'].post
  update:
    x-apievangelist-phrasing:
      intent: Start a browser agent task
      effect: write
      questions:
      - How do I tell an AI agent to go do something in a browser for me?
      - Can I limit which domains the agent visits and how many steps it takes?
      - Is it possible to get the task result back as structured JSON matching my schema?
      - Can I run a follow-up task in a session I already have open?
      instructions:
      - text: Create a task to {task}, starting at {start_url}.
        slots:
          task: requestBody.task
          start_url: requestBody.startUrl
      - text: Run {task} in existing session {session_id}.
        slots:
          task: requestBody.task
          session_id: requestBody.sessionId
      - text: Have the agent {task}, staying on {domains} and stopping after {max_steps} steps.
        slots:
          task: requestBody.task
          domains: requestBody.allowedDomains
          max_steps: requestBody.maxSteps
      - text: Start a task to {task} and return output matching schema {schema}.
        slots:
          task: requestBody.task
          schema: requestBody.structuredOutput
      method: generated
      generated: '2026-09-26'
- target: $.paths['/tasks/{task_id}'].get
  update:
    x-apievangelist-phrasing:
      intent: Get a task's full details, steps and files
      effect: read
      questions:
      - How do I see every step the agent took on a task and the files it produced?
      - Where can I get the complete record of a task once it has finished?
      instructions:
      - text: Show the full details, steps and output files of task {task_id}.
        slots:
          task_id: path.task_id
      - text: Walk me through what the agent did step by step in task {task_id}.
        slots:
          task_id: path.task_id
      method: generated
      generated: '2026-09-26'
- target: $.paths['/tasks/{task_id}'].patch
  update:
    x-apievangelist-phrasing:
      intent: Pause, resume or stop a task
      effect: write
      questions:
      - Can I pause an agent task partway through and resume it later?
      - How do I stop a task and also shut down the session it's running in?
      instructions:
      - text: Pause task {task_id}.
        slots:
          task_id: path.task_id
      - text: Apply action {action} to task {task_id}.
        slots:
          action: requestBody.action
          task_id: path.task_id
      - text: Resume the paused task {task_id}.
        slots:
          task_id: path.task_id
      method: generated
      generated: '2026-09-26'
- target: $.paths['/tasks/{task_id}/status'].get
  update:
    x-apievangelist-phrasing:
      intent: Poll a task's status, output and cost
      effect: read
      questions:
      - What's the most efficient way to poll whether my agent task is done?
      - How much has a task cost so far, and what did it return?
      instructions:
      - text: Poll task {task_id} until it finishes and give me its output.
        slots:
          task_id: path.task_id
      - text: Check only the status and cost of task {task_id}.
        slots:
          task_id: path.task_id
      method: generated
      generated: '2026-09-26'
- target: $.paths['/tasks/{task_id}/logs'].get
  update:
    x-apievangelist-phrasing:
      intent: Download a task's execution logs
      effect: read
      questions:
      - Where can I download the detailed execution logs for a task?
      - How do I debug a failed agent task from its logs?
      instructions:
      - text: Get the log download link for task {task_id}.
        slots:
          task_id: path.task_id
      - text: Fetch the step-by-step execution logs of task {task_id} so I can debug it.
        slots:
          task_id: path.task_id
      method: generated
      generated: '2026-09-26'