---
title: "Schedule Sim Run"
method: POST
path: "/api/v1/sim-runs"
tags: ["simulated runs"]
---

# Schedule Sim Run

`POST /api/v1/sim-runs`

## Request

- Base URL: `https://api.brumby.com` (the document declares no server; this is the origin it was published from)
- URL: `https://api.brumby.com/api/v1/sim-runs`
- Auth: HTTP bearer

## Request body

- SimRunCreate: Schedule one simulated run. `mode` and `destination` are what a farmer's request carries (brumby-cloud `buildMissionConfig`): a survey of one paddock, or a muster from `paddock` to `destination`. They are composed into the run's mission config in the dock's own shape -- `mode`, and `paddocks` as `[source, destination]` -- so there is no second representation to drift. Omitted, the run is what it always was: the composer's muster of one paddock.
  - `weather` SimWeather
    - `wind` SimWind
      - `speed_ms` number
      - `from_deg` number
      - `vertical_ms` number
      - `gust` SimGust
        - `enabled` boolean
        - `magnitude_sigma_ms` number
        - `direction_sigma_deg` number
        - `magnitude_time_constant_s` number
        - `direction_time_constant_s` number
        - `peak_sigma` number
    - `air` SimAir
      - `temperature_c` number
      - `humidity_pct` number
      - `rain` boolean
      - `visibility_m` number
      - `cloud_base_ft` number
    - `time_of_day_h` number
  - `farm_id` string, required
  - `device_id` string, required
  - `paddock` string, nullable
  - `mode` 'survey' | 'muster' | 'flight_plan', nullable
  - `destination` string, nullable
  - `flight_plan` object, nullable
  - `herd` SimHerd: How many head, placed how, for one run. Every field is optional and an omitted one means "whatever the sim's own cattle sheet says". That matters: a sweep over seeds should not have to restate the mob profiles to leave them alone, and a run with an empty herd spec is exactly the run `pixi run sim` gives you today.
    - `seed` integer, nullable
    - `cows` integer, nullable
    - `clusters` integer, nullable
    - `cluster_spread_m` number, nullable
    - `stragglers` integer, nullable
    - `profile_mix` object, nullable
  - `dispatch` 'launch' | 'headless': How a simulated run reaches an aircraft, and therefore who authorises it. LAUNCH is the ordinary path and the only one that flies anything: the run schedules a mission request exactly as the customer app does, a pilot signs the JSA and presses LAUNCH, and the dock delivers. FleetOps adds nothing to that sequence and skips nothing in it. HEADLESS never asks a dock for anything. The bench composes the same mission config from the same farm and runs it against a stub dock, so there is no clearance to grant, no aircraft to arm and nothing to authorise -- which is why there is no JSA on it. A sweep of two hundred seeds is this one; a run you want a flight record for is the other.
  - `autonomy_host` 'bench' | 'jetson'
  - `mission_id` string, nullable
  - `mission_plan_id` string, uuid, nullable
  - `mission_plan_revision` integer, nullable

## Response `201`

Successful Response

- SimRunResponse: One cell of a sweep, as FleetOps holds it.
  - `run_id` string, uuid, required
  - `batch_id` string, uuid, nullable, required
  - `ordinal` integer, required
  - `farm_id` string, required
  - `device_id` string, required
  - `world_slug` string, required
  - `paddock` string, nullable, required
  - `autonomy_host` 'bench' | 'jetson'
  - `simulator_host` string, nullable
  - `runner_launch_mode` 'worker' | 'command'
  - `weather` object
  - `mission_config` object, nullable
  - `mission_plan_id` string, uuid, nullable
  - `mission_plan_revision` integer, nullable
  - `claim_token` string, uuid, nullable
  - `lease_expires_at` string, date-time, nullable
  - `cancel_requested` boolean
  - `herd` object, required
  - `dispatch` 'launch' | 'headless', required: How a simulated run reaches an aircraft, and therefore who authorises it. LAUNCH is the ordinary path and the only one that flies anything: the run schedules a mission request exactly as the customer app does, a pilot signs the JSA and presses LAUNCH, and the dock delivers. FleetOps adds nothing to that sequence and skips nothing in it. HEADLESS never asks a dock for anything. The bench composes the same mission config from the same farm and runs it against a stub dock, so there is no clearance to grant, no aircraft to arm and nothing to authorise -- which is why there is no JSA on it. A sweep of two hundred seeds is this one; a run you want a flight record for is the other.
  - `state` 'pending' | 'requested' | 'claimed' | 'running' | 'succeeded' | 'failed' | 'cancelled', required: Where one cell of a sweep has got to. PENDING through FINISHED is the bench's own progress. REQUESTED is where a LAUNCH run waits: its mission request exists and the pilot's queue owns it from there, so FleetOps deliberately stops describing it -- the request's own state is the answer, and a copy here would be a second one.
  - `mission_request_id` string, uuid, nullable, required
  - `simulation_run_id` string, uuid, nullable, required
  - `claimed_by` string, nullable, required
  - `claimed_at` string, date-time, nullable, required
  - `started_at` string, date-time, nullable, required
  - `finished_at` string, date-time, nullable, required
  - `outcome` object, nullable, required
  - `failure` string, nullable, required
  - `created_by` string, required
  - `created_at` string, date-time, required

## Other responses

- `422`: Validation Error

---

[API](https://skmtc.dev/brumby/apis/brumby-platform-api.md) - [All operations](https://skmtc.dev/brumby/apis/brumby-platform-api/llms.txt) - [OpenAPI document](https://skmtc.dev/brumby/apis/brumby-platform-api/revisions/fe185e20813e?raw)