Skip to main content
Auto-label jobs use trained models to automatically generate predictions and labels for unlabeled data in your projects.

​
Methods

​
list

List all auto-label jobs. 
client.auto_label_jobs.list(
    project="proj_123",
    limit=100,
    cursor="next_page_token"
)
​
project
str
Filter jobs by project UID
​
limit
int
Maximum number of jobs to return per page
​
cursor
str
Pagination cursor for fetching the next page
​
data
list[AutoLabelJob]
List of auto-label job objects
​
next_cursor
str
Cursor for the next page of results

​
get

Retrieve a specific auto-label job by UID. 
client.auto_label_jobs.get(uid="job_123")
​
uid
str
required
Unique identifier of the auto-label job
​
uid
str
Unique identifier of the job
​
project
str
Project UID where the job is running
​
model_type
str
Type of model used for labeling
​
confidence_threshold
float
Minimum confidence score for predictions
​
labels
list[str]
Labels to generate predictions for
​
status
str
Current status of the job (pending, running, completed, failed)
​
created_at
str
Timestamp when the job was created
​
completed_at
str
Timestamp when the job completed

​
create

Create a new auto-label job for a project. 
client.auto_label_jobs.create(
    project_uid="proj_123",
    model_type="yolov8",
    confidence_threshold=0.8,
    labels=["person", "car", "bicycle"],
    dry_run=False
)
​
project_uid
str
required
UID of the project to run auto-labeling on
​
model_type
str
Type of model to use (e.g., “yolov8”, “faster-rcnn”)
​
confidence_threshold
float
Minimum confidence score for predictions (0.0 to 1.0)
​
labels
list[str]
Specific labels to generate predictions for. If not provided, uses all project labels.
​
dry_run
bool
If true, preview results without saving labels (default: false)
​
uid
str
Unique identifier of the created job
​
status
str
Initial status of the job

​
cancel

Cancel a running auto-label job. 
client.auto_label_jobs.cancel(uid="job_123")
​
uid
str
required
Unique identifier of the job to cancel

​
Async Methods

All methods are available in async form through AsyncAvala: 
await client.auto_label_jobs.list(project="proj_123")
await client.auto_label_jobs.get(uid="job_123")
await client.auto_label_jobs.create(
    project_uid="proj_123",
    model_type="yolov8"
)
await client.auto_label_jobs.cancel(uid="job_123")

Build docs developers (and LLMs) love

Get started for freeTalk to us