Execute Orchestration

Start an AI orchestration execution with full provenance tracking and step-level auditing.

Initiates an AI orchestration execution. The orchestration engine tracks every step, LLM inference call, tool invocation, and decision point — producing a complete, transparency-log-anchored audit trail.

Orchestrations are the primary unit of AI agent work in MAIP. They capture the full lifecycle of an agent’s reasoning and actions, enabling compliance teams to audit what the agent did, why it did it, and what data it consumed.

Authentication

X-API-Key · string · required

API key with orchestrations:write scope. Alternatively, pass a Bearer JWT token in the Authorization header.

X-Tenant-ID · string · required

Tenant identifier for multi-tenant isolation.

Request

agent_id · string · required

MAIP agent identifier executing the orchestration.

workflow_id · string

Optional workflow definition identifier. If provided, the orchestration follows the workflow’s DAG of steps.

input · object · required

Input data for the orchestration. Schema depends on the workflow definition or agent’s expected input format.

model · string

Default LLM model identifier to use for inference steps (e.g. claude-sonnet-4-20250514, gpt-4o). Individual steps may override this.

max_steps · integer

Maximum number of steps the orchestration may execute before automatic termination. Defaults to 10.

timeout_seconds · integer

Maximum wall-clock duration in seconds before the orchestration is timed out. Defaults to 300 (5 minutes).

Response

id · string

Unique orchestration identifier in MAIP format (maip-orch:ULID).

agent_id · string

The agent executing the orchestration.

workflow_id · string

The workflow definition identifier, if provided.

status · string

Execution status. Always running on creation.

started_at · string

ISO 8601 timestamp when execution began.