Skip to main content
Agents are automation workflows that trigger callbacks when specific labeling events occur. They can monitor task completion, quality changes, and other project events.

​
Methods

​
list

List all agents. 
client.agents.list(
    limit=100,
    cursor="next_page_token"
)
​
limit
int
Maximum number of agents to return per page
​
cursor
str
Pagination cursor for fetching the next page
​
data
list[Agent]
List of agent objects
​
next_cursor
str
Cursor for the next page of results

​
get

Retrieve a specific agent by UID. 
client.agents.get(uid="agent_123")
​
uid
str
required
Unique identifier of the agent
​
uid
str
Unique identifier of the agent
​
name
str
Name of the agent
​
events
list[str]
List of events that trigger this agent
​
description
str
Description of the agent’s purpose
​
callback_url
str
URL to call when events are triggered
​
is_active
bool
Whether the agent is currently active
​
project
str
Project UID this agent is associated with
​
task_types
list[str]
Types of tasks this agent monitors

​
create

Create a new agent. 
client.agents.create(
    name="Quality Monitor",
    events=["task.completed", "task.reviewed"],
    description="Monitor task completion and quality",
    callback_url="https://api.example.com/webhook",
    is_active=True,
    project="proj_123",
    task_types=["classification", "segmentation"]
)
​
name
str
required
Name of the agent
​
events
list[str]
List of events that trigger this agent (e.g., “task.completed”, “task.reviewed”)
​
description
str
Description of the agent’s purpose
​
callback_url
str
URL to call when events are triggered
​
is_active
bool
Whether the agent should be active immediately (default: true)
​
project
str
Project UID to associate this agent with
​
task_types
list[str]
Types of tasks this agent should monitor

​
update

Update an existing agent. 
client.agents.update(
    uid="agent_123",
    name="Updated Quality Monitor",
    is_active=False
)
​
uid
str
required
Unique identifier of the agent to update
​
name
str
New name for the agent
​
events
list[str]
Updated list of events that trigger this agent
​
description
str
Updated description
​
callback_url
str
Updated callback URL
​
is_active
bool
Whether the agent should be active
​
project
str
Updated project UID
​
task_types
list[str]
Updated task types to monitor

​
delete

Delete an agent. 
client.agents.delete(uid="agent_123")
​
uid
str
required
Unique identifier of the agent to delete

​
list_executions

List execution history for an agent. 
client.agents.list_executions(
    uid="agent_123",
    limit=50,
    cursor="next_page_token"
)
​
uid
str
required
Unique identifier of the agent
​
limit
int
Maximum number of executions to return per page
​
cursor
str
Pagination cursor for fetching the next page
​
data
list[AgentExecution]
List of agent execution records
​
next_cursor
str
Cursor for the next page of results

​
test

Test an agent by triggering a test execution. 
client.agents.test(uid="agent_123")
​
uid
str
required
Unique identifier of the agent to test
​
status
str
Status of the test execution
​
response
dict
Response from the callback URL

​
Async Methods

All methods are available in async form through AsyncAvala: 
await client.agents.list()
await client.agents.get(uid="agent_123")
await client.agents.create(name="Quality Monitor")
await client.agents.update(uid="agent_123", is_active=False)
await client.agents.delete(uid="agent_123")
await client.agents.list_executions(uid="agent_123")
await client.agents.test(uid="agent_123")

Build docs developers (and LLMs) love

Get started for freeTalk to us