Schemas

Always list for paginated list responses.

Goal interfaces in stable opaque-id order. Do not assume the first item is the Goal your workflow should execute; store the intended goal_id when it is published.

Opaque cursor for the next page. null means there are no more Goals.

Always goal for Goal responses.

Opaque Goal identity. This is distinct from the nested RunSpec identity.

Short developer-facing title from the current published RunSpec. It may be null for an untitled Goal.

Developer-facing summary of what the current published Goal does. This is not the execution prompt.

Always active; non-executable Goals are not returned by this surface.

Currently published immutable interface used to validate new variables and interpret result. New business keys pin this version at acceptance time.

Opaque immutable RunSpec identity.

Monotonic published version within this Goal. A later publish affects only new Runs.

Normalized JSON Schema for per-Run variables. Respect required, property types, enum values, defaults, and additionalProperties; invalid input is rejected before execution.

Normalized JSON Schema for result. CALL-E exposes a result only after it is validated against this exact pinned schema and durably persisted.

Recipient phone in canonical E.164 form: +, country code, and subscriber number with no spaces, punctuation, or extension. The caller must be authorized to contact it. CALL-E validates it against the published Goal's fixed Voice Target policy.

Per-Run business context validated against the pinned published input_schema. Keys and required fields vary by Goal. Values must be finite JSON strings, numbers, or booleans; nested objects, arrays, and null are not supported. Omit for Goals whose schema accepts {}.

Public Goal Run identity. Persist this value and use it as goal_run_id when polling.

Goal identity supplied in the create path.

Internal execution member exposed for correlation; do not use it in the Goal Run polling path.

Read-only identity and version of the exact RunSpec pinned by this Run.

Stable telephone execution state. queued and in_progress are non-terminal; completed, failed, and canceled are terminal. A completed call can still have result: null and error: null briefly while CALL-E parses and saves the result.

Parsed result validated against the published result schema and durably persisted, or null while processing or when the Run has an error. Its keys vary by Goal.

Unified execution or result-processing error, or null. Branch on code; keep message for logs and operators. A non-null error is final and is mutually exclusive with result.

UTC time at which CALL-E durably accepted this Goal Run.

UTC telephone-execution completion time, or null while execution is non-terminal.

Exact immutable RunSpec id pinned when the Goal Run was accepted.

Published RunSpec version pinned for this Goal Run.

Human-readable safe explanation. Do not parse this field for application logic.

Optional low-cardinality diagnostic detail safe for logs or narrow application branching.

Natural-language instruction for the call task. Include the goal, relevant details the voice agent should know, and the exact information you want collected.

Optional explicit recipients for this call task. Omit it when the task text already contains the phone targets CALL-E should use.

Optional JSON Schema object that defines the structured result CALL-E should extract for the whole call task.

CALL-E passes the schema, including field description values, to the extraction model after the call reaches a terminal state. Use descriptions to explain field meaning and enum selection logic, for example: "Use strong when the prospect asks about pricing, demos, or next steps."

Descriptions guide extraction but are not hard validation rules. Hard validation comes from type, required, enum, and additionalProperties.

Supported schema features are type, properties, required, enum, nested object fields, simple array.items, description, and additionalProperties: false. Unsupported features include $ref, oneOf, anyOf, allOf, recursive schemas, complex format validation, and additionalProperties: true.

Prefer string enums over booleans for business decisions that may be unclear, and include an unknown enum value when the call may not provide enough evidence.

Optional JSON Schema object that defines the structured result CALL-E should extract independently for each recipient.

This is useful for batch calls where each recipient needs their own outcome, such as can_attend, confirmed, requested_callback, or interest_level.

Do not use reserved recipient response field names such as summary, status, transcript, call_id, or timing fields as custom result fields. Use names such as customer_summary, notes, or reason instead.

Field description values are passed to the extraction model and should explain how enum values should be selected. Descriptions guide extraction but are not hard validation rules. Hard validation comes from type, required, enum, and additionalProperties.

Object schemas are strict by default. Fields not declared in properties are rejected, and unsupported or invalid recipient results are returned as null.

Optional caller-owned metadata echoed on the call and webhook payloads. Use this for workflow ids, tenant ids, or internal correlation keys.

Optional per-request HTTPS webhook URL. When provided, CALL-E sends terminal call events to this URL in addition to project-level webhook delivery.

Phone numbers in E.164 format for this recipient. Replace placeholders such as <RECIPIENT_1_E164_PHONE> with phone numbers you own or are authorized to call.

BCP 47 locale hint for the conversation, for example en-US.

Recipient country or region code used for routing and compliance checks, for example US.

Confidence score in the 0 to 1 range for CALL-E's task completion judgment.

Confidence label for the task completion judgment, for example low, medium, or high.

Seconds from the start of the attempt. null when the source line did not include a parseable timestamp.

Speaker for this transcript turn.

Spoken text for this transcript turn.

Stable outbound attempt identifier.

Phone number dialed by this attempt.

Current attempt lifecycle state.

ISO 8601 timestamp when this attempt started dialing. null before dialing begins.

ISO 8601 timestamp when this attempt reached a terminal state. null while queued, dialing, or in progress.

Human-readable summary for this attempt. null when no attempt-level summary is available.

Structured transcript turns for this attempt. Empty when no transcript is available.

Provider call identifier for support correlation when available.

Machine-readable failure reason when this attempt failed.

Human-readable failure explanation when this attempt failed.

Stable recipient identifier within the call task.

Phone numbers associated with this recipient.

BCP 47 locale hint used for this recipient when available.

Country or region code used for routing and compliance checks when available.

Current recipient lifecycle state.

Schema-valid structured result object extracted for this recipient using recipient_result_schema.

null means CALL-E could not produce a schema-valid result for this recipient from the terminal call evidence, or no recipient_result_schema was provided.

Short human-readable summary for this recipient. null while the recipient is still running or when no useful summary is available.

Outbound dial attempts made for this recipient.

Public CALL-E call task identifier. Store this id to fetch state, list events, and correlate webhooks.

Always call_task for call task responses.

Current call task lifecycle state.

Original task instruction submitted when the call task was created.

Recipient states for this call task.

Schema-valid structured result object extracted for the whole call task using result_schema.

null means CALL-E could not produce a schema-valid task-level result from the terminal call evidence, or no result_schema was provided. Check recipient-level structured_result when you use recipient_result_schema for batch calls.

Short human-readable summary of the call task outcome. null while the call task is still running or when no useful summary is available.

Post-summary judgment for whether the task reached a clear end state for the user. null until CALL-E has a terminal post-summary outcome.

Confidence for task_completed. null until CALL-E has a terminal post-summary outcome.

Short evidence items supporting the post-summary task outcome. Empty until CALL-E has terminal evidence.

Caller-owned metadata submitted on the create call request.

Machine-readable failure reason when status is failed; otherwise null.

Human-readable failure explanation when status is failed; otherwise null.

ISO 8601 timestamp when CALL-E accepted the call request.

ISO 8601 timestamp when the call reached a terminal state. null while queued or in progress.

Public event identifier.

Developer-facing event type, for example call.completed.

Public CALL-E call identifier associated with this event.

ISO 8601 timestamp when the event was emitted.

Event severity for log routing and alerting.

Call status at the time this event was emitted.

Short human-readable event message.

Event-specific structured details. Shape depends on the event type.

Always list for paginated list responses.

Events in this page, ordered from oldest to newest for the requested cursor window.

Cursor for the next page. null means there are no more events.

Unique webhook event id. Use it as an idempotency key when processing webhook side effects.

Terminal event type for the call.

ISO 8601 timestamp when CALL-E created this webhook event.

Terminal call task snapshot associated with the webhook event.

Public CALL-E call task identifier. Store this id to fetch state, list events, and correlate webhooks.

Always call_task for call task responses.

Current call task lifecycle state.

Original task instruction submitted when the call task was created.

Recipient states for this call task.

Schema-valid structured result object extracted for the whole call task using result_schema.

null means CALL-E could not produce a schema-valid task-level result from the terminal call evidence, or no result_schema was provided. Check recipient-level structured_result when you use recipient_result_schema for batch calls.

Short human-readable summary of the call task outcome. null while the call task is still running or when no useful summary is available.

Post-summary judgment for whether the task reached a clear end state for the user. null until CALL-E has a terminal post-summary outcome.

Confidence for task_completed. null until CALL-E has a terminal post-summary outcome.

Short evidence items supporting the post-summary task outcome. Empty until CALL-E has terminal evidence.

Caller-owned metadata submitted on the create call request.

Machine-readable failure reason when status is failed; otherwise null.

Human-readable failure explanation when status is failed; otherwise null.

ISO 8601 timestamp when CALL-E accepted the call request.

ISO 8601 timestamp when the call reached a terminal state. null while queued or in progress.

Optional acknowledgement flag returned by your server.