Skip to content

automations

Automation models for the Open WebUI scheduling system.

Automations are scheduled tasks that execute prompts on a recurring basis using RRULE scheduling. Each automation creates a chat and runs through the full chat completion pipeline when triggered.

Classes

AutomationTerminalConfig

Bases: BaseModel

Configuration for a terminal server attached to an automation.

When present, the automation's execution will connect to the specified terminal server before running the prompt.

Returned by:

Attributes

server_id 
server_id: str

ID of the terminal server to connect to.

cwd 
cwd: Optional[str] = None

Working directory to set on the terminal server before execution.

AutomationData

Bases: BaseModel

Core data payload defining what an automation does and when it runs.

This is the structured form used in create/update requests. When stored in the database and returned by the API, it is serialized as a plain dict on AutomationModel.data.

Returned by:

Attributes

prompt 
prompt: str

Prompt template to execute. Supports template variables like {{user_name}}.

model_id 
model_id: str

ID of the model to use for chat completion.

rrule 
rrule: str

RRULE recurrence rule defining the schedule (RFC 5545). For example, FREQ=DAILY;DTSTART=20250101T090000Z for daily at 9 AM UTC.

terminal 
terminal: Optional[AutomationTerminalConfig] = None

Optional terminal server configuration.

AutomationModel

Bases: BaseModel

Full automation model as stored in the database.

Represents an automation with all persisted fields including scheduling state.

Subclassed by:

Attributes

id 
id: str

Unique identifier (UUID).

user_id 
user_id: str

ID of the user who owns the automation.

name 
name: str

Display name of the automation.

data 
data: dict[str, Any]

Serialized automation data defining the prompt, schedule, and model.

Dict Fields
  • prompt (str, required): Prompt template to execute
  • model_id (str, required): Model ID for chat completion
  • rrule (str, required): RRULE recurrence rule for scheduling
  • terminal (dict, optional): Terminal server config with server_id and optional cwd keys
meta 
meta: Optional[dict[str, Any]] = None

Optional metadata for the automation.

Dict Fields
  • system_prompt (str, optional): System prompt injected before the user prompt
  • temperature (float, optional): Sampling temperature override
  • max_tokens (int, optional): Maximum tokens override
  • webhook (str, optional): Webhook URL notified after execution
is_active 
is_active: bool = True

Whether the automation is active and eligible for scheduling.

last_run_at 
last_run_at: Optional[int] = None

Timestamp (nanoseconds since epoch) of the last execution.

next_run_at 
next_run_at: Optional[int] = None

Timestamp (nanoseconds since epoch) of the next scheduled execution.

created_at 
created_at: int

Timestamp (nanoseconds since epoch) of creation.

updated_at 
updated_at: int

Timestamp (nanoseconds since epoch) of last update.

AutomationRunModel

Bases: BaseModel

Record of a single automation execution.

Each time an automation runs, a run record is created tracking the outcome.

Returned by:

Attributes

id 
id: str

Unique identifier (UUID).

automation_id 
automation_id: str

ID of the parent automation.

chat_id 
chat_id: Optional[str] = None

ID of the chat created by this run, if successful.

status 
status: str

Execution status: 'success' or 'error'.

error 
error: Optional[str] = None

Error message if status is 'error'.

created_at 
created_at: int

Timestamp (nanoseconds since epoch) of when the run completed.

AutomationForm

Bases: BaseModel

Form for creating or updating an automation.

The data field contains the structured automation definition including the RRULE schedule. The backend validates the RRULE against the user's timezone and enforces rate limits.

Used by:

Attributes

name 
name: str

Display name of the automation.

data 
data: AutomationData

Core automation data: prompt, model, schedule, and optional terminal config.

meta 
meta: Optional[dict[str, Any]] = None

Optional metadata for the automation.

Dict Fields
  • system_prompt (str, optional): System prompt injected before the user prompt
  • temperature (float, optional): Sampling temperature override
  • max_tokens (int, optional): Maximum tokens override
  • webhook (str, optional): Webhook URL notified after execution
is_active 
is_active: Optional[bool] = True

Whether the automation should be active on creation. Defaults to True.

AutomationResponse

Bases: AutomationModel

Enriched automation response with run history and computed next runs.

Extends AutomationModel with the latest run record and a list of upcoming execution timestamps.

Returned by:

Attributes

last_run 
last_run: Optional[AutomationRunModel] = None

The most recent run record, if any.

next_runs 
next_runs: Optional[list[int]] = None

List of upcoming execution timestamps (nanoseconds since epoch).

AutomationListResponse

Bases: BaseModel

Paginated list of automations with enriched run data.

Each item includes the latest run record. The total field reflects the total count before pagination.

Returned by:

Attributes

items 
items: list[AutomationResponse] = []

List of automations in the current page.

total 
total: int = 0

Total number of automations matching the query.