Python SDK Reference - Avala Documentation

Auto-generated from SDK source. Last updated: February 2026.

Client

`avala.Client`

Synchronous client for the Avala API.

from avala import Client

client = Client(
    api_key="avk_...",       # or set AVALA_API_KEY env var
    base_url="https://api.avala.ai/api/v1",  # default
    timeout=30.0,            # seconds
    max_retries=2,
)

Parameter	Type	Default	Description
`api_key`	`str \| None`	`AVALA_API_KEY` env var	Your Avala API key.
`base_url`	`str \| None`	`https://api.avala.ai/api/v1`	API base URL.
`timeout`	`float`	`30.0`	Request timeout in seconds.
`max_retries`	`int`	`2`	Automatic retries on transient failures (5xx, timeouts).

Resources:

Property	Type	Description
`client.datasets`	`Datasets`	Dataset operations
`client.projects`	`Projects`	Project operations (read-only)
`client.exports`	`Exports`	Export operations
`client.tasks`	`Tasks`	Task operations (read-only)
`client.storage_configs`	`StorageConfigs`	Cloud storage operations
`client.agents`	`Agents`	Automation agent operations
`client.inference_providers`	`InferenceProviders`	Inference provider operations
`client.auto_label_jobs`	`AutoLabelJobs`	Auto-labeling job operations
`client.quality_targets`	`QualityTargets`	Quality target operations (project-scoped)
`client.consensus`	`Consensus`	Consensus scoring operations (project-scoped)
`client.webhooks`	`Webhooks`	Webhook subscription operations
`client.webhook_deliveries`	`WebhookDeliveries`	Webhook delivery log operations (read-only)
`client.organizations`	`Organizations`	Organization management
`client.slices`	`Slices`	Slice operations

Methods:

Method	Returns	Description
`client.rate_limit_info`	`dict[str, str \| None]`	Rate limit headers from last response
`client.close()`	`None`	Close the HTTP transport

Supports context manager:

with Client() as client:
    datasets = client.datasets.list()

`avala.AsyncClient`

Asynchronous client with the same interface. All resource methods are async.

from avala import AsyncClient

async with AsyncClient() as client:
    datasets = await client.datasets.list()

Resources

datasets

`client.datasets.create()`

Create a new dataset.

dataset = client.datasets.create(
    name="LiDAR Captures Q1",
    slug="lidar-captures-q1",
    data_type="lidar",
    is_sequence=True,
    visibility="private",
    provider_config={
        "provider": "aws_s3",
        "s3_bucket_name": "my-datasets",
        "s3_bucket_region": "us-east-1",
        "s3_bucket_prefix": "captures/q1/",
        "s3_access_key_id": "AKIA...",
        "s3_secret_access_key": "your-secret-key",
    },
    owner_name="my-org",
)

Parameter	Type	Default	Description
`name`	`str`	—	Display name for the dataset.
`slug`	`str`	—	URL-friendly identifier.
`data_type`	`str`	—	Data type: `image`, `video`, `lidar`, or `mcap`.
`is_sequence`	`bool`	`False`	Whether the dataset contains sequences.
`visibility`	`str`	`"private"`	`"private"` or `"public"`.
`create_metadata`	`bool`	`True`	Whether to create dataset metadata.
`provider_config`	`dict \| None`	`None`	Cloud storage provider configuration.
`owner_name`	`str \| None`	`None`	Dataset owner username or email.

Returns: Dataset

`client.datasets.list()`

List datasets with optional filtering and cursor-based pagination.

page = client.datasets.list(data_type="image", name="highway", limit=20)

Parameter	Type	Default	Description
`data_type`	`str \| None`	`None`	Filter by data type: `image`, `video`, `lidar`, `mcap`, or `splat`.
`name`	`str \| None`	`None`	Filter by name (case-insensitive substring match).
`status`	`str \| None`	`None`	Filter by status: `creating` or `created`.
`visibility`	`str \| None`	`None`	Filter by visibility: `private` or `public`.
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor from a previous response.

Returns: CursorPage[Dataset]

`client.datasets.get()`

Get a single dataset by UID.

dataset = client.datasets.get("550e8400-e29b-41d4-a716-446655440000")

Parameter	Type	Description
`uid`	`str`	Dataset unique identifier.

Returns: Dataset

`client.datasets.list_items()`

List items in a dataset. Uses the owner/slug path.

page = client.datasets.list_items("my-org", "my-dataset", limit=50)

Parameter	Type	Default	Description
`owner`	`str`	—	Dataset owner (organization slug).
`slug`	`str`	—	Dataset slug.
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[DatasetItem]

`client.datasets.get_item()`

Get a single item within a dataset.

item = client.datasets.get_item("my-org", "my-dataset", "item_uid")

Parameter	Type	Description
`owner`	`str`	Dataset owner (organization slug).
`slug`	`str`	Dataset slug.
`item_uid`	`str`	Item unique identifier.

Returns: DatasetItem

`client.datasets.list_sequences()`

List sequences in a dataset.

page = client.datasets.list_sequences("my-org", "my-dataset", limit=50)

Parameter	Type	Default	Description
`owner`	`str`	—	Dataset owner (organization slug).
`slug`	`str`	—	Dataset slug.
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[DatasetSequence]

`client.datasets.get_sequence()`

Get a single sequence within a dataset.

seq = client.datasets.get_sequence("my-org", "my-dataset", "sequence_uid")

Parameter	Type	Description
`owner`	`str`	Dataset owner (organization slug).
`slug`	`str`	Dataset slug.
`sequence_uid`	`str`	Sequence unique identifier.

Returns: DatasetSequence

projects

`client.projects.list()`

List projects with cursor-based pagination.

page = client.projects.list(limit=20)

Parameter	Type	Default	Description
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[Project]

`client.projects.get()`

Get a single project by UID.

project = client.projects.get("770a9600-a40d-63f6-c938-668877660000")

Parameter	Type	Description
`uid`	`str`	Project unique identifier.

Returns: Project

tasks

`client.tasks.list()`

List tasks with optional filtering and cursor-based pagination.

page = client.tasks.list(project="proj_uid", status="pending", limit=50)

Parameter	Type	Default	Description
`project`	`str \| None`	`None`	Filter by project UID.
`status`	`str \| None`	`None`	Filter by task status.
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[Task]

`client.tasks.get()`

Get a single task by UID.

task = client.tasks.get("990c1800-b62f-85a8-e150-880099880000")

Parameter	Type	Description
`uid`	`str`	Task unique identifier.

Returns: Task

exports

`client.exports.list()`

List exports with cursor-based pagination.

page = client.exports.list(limit=10)

Parameter	Type	Default	Description
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[Export]

`client.exports.create()`

Create a new export. Provide either project or dataset.

export = client.exports.create(project="proj_uid")

Parameter	Type	Default	Description
`project`	`str \| None`	`None`	Project UID to export.
`dataset`	`str \| None`	`None`	Dataset UID to export.

Returns: Export

`client.exports.get()`

Get a single export by UID.

export = client.exports.get("export_uid")

Parameter	Type	Description
`uid`	`str`	Export unique identifier.

Returns: Export

storage_configs

`client.storage_configs.list()`

List storage configurations.

page = client.storage_configs.list()

Parameter	Type	Default	Description
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[StorageConfig]

`client.storage_configs.create()`

Create a new storage configuration.

import os

config = client.storage_configs.create(
    name="Production S3",
    provider="s3",
    s3_bucket_name="my-bucket",
    s3_bucket_region="us-west-2",
    s3_access_key_id=os.environ["AWS_ACCESS_KEY_ID"],
    s3_secret_access_key=os.environ["AWS_SECRET_ACCESS_KEY"],
)

Parameter	Type	Required	Description
`name`	`str`	Yes	Configuration name.
`provider`	`str`	Yes	`"s3"` or `"google_cloud"`.
`s3_bucket_name`	`str \| None`	No	S3 bucket name.
`s3_bucket_region`	`str \| None`	No	S3 region (e.g., `us-west-2`).
`s3_bucket_prefix`	`str \| None`	No	S3 key prefix.
`s3_access_key_id`	`str \| None`	No	AWS access key ID.
`s3_secret_access_key`	`str \| None`	No	AWS secret access key.
`s3_is_accelerated`	`bool \| None`	No	Enable S3 Transfer Acceleration.
`gc_storage_bucket_name`	`str \| None`	No	GCS bucket name.
`gc_storage_prefix`	`str \| None`	No	GCS key prefix.
`gc_storage_auth_json_content`	`str \| None`	No	GCS service account JSON.

Returns: StorageConfig

`client.storage_configs.test()`

Test a storage configuration.

result = client.storage_configs.test("config_uid")

Returns: dict[str, Any]

`client.storage_configs.delete()`

Delete a storage configuration.

client.storage_configs.delete("config_uid")

Returns: None

agents

`client.agents.list()`

List automation agents.

page = client.agents.list(limit=20)

Parameter	Type	Default	Description
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[Agent]

`client.agents.get()`

Get a single agent by UID.

agent = client.agents.get("agent_uid")

Parameter	Type	Description
`uid`	`str`	Agent unique identifier.

Returns: Agent

`client.agents.create()`

Create a new automation agent.

agent = client.agents.create(
    name="QA Bot",
    events=["task.completed"],
    callback_url="https://example.com/webhook",
    task_types=["annotation"],
)

Parameter	Type	Required	Description
`name`	`str`	Yes	Agent name.
`events`	`list[str] \| None`	No	Events the agent listens to.
`description`	`str \| None`	No	Agent description.
`callback_url`	`str \| None`	No	URL called when events fire.
`is_active`	`bool \| None`	No	Enable or disable the agent.
`project`	`str \| None`	No	Scope to a specific project UID.
`task_types`	`list[str] \| None`	No	Task types the agent handles.

Returns: Agent

`client.agents.update()`

Update an existing agent.

agent = client.agents.update("agent_uid", is_active=False, callback_url="https://example.com/new-webhook")

Parameter	Type	Required	Description
`uid`	`str`	Yes	Agent unique identifier.
`name`	`str \| None`	No	Agent name.
`events`	`list[str] \| None`	No	Events to listen to.
`description`	`str \| None`	No	Agent description.
`callback_url`	`str \| None`	No	Callback URL.
`is_active`	`bool \| None`	No	Enable or disable the agent.
`project`	`str \| None`	No	Project UID scope.
`task_types`	`list[str] \| None`	No	Task types.

Returns: Agent

`client.agents.delete()`

Delete an agent.

client.agents.delete("agent_uid")

Returns: None

`client.agents.list_executions()`

List execution history for an agent.

page = client.agents.list_executions("agent_uid", limit=50)

Parameter	Type	Default	Description
`uid`	`str`	—	Agent unique identifier.
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[AgentExecution]

`client.agents.test()`

Send a test event to an agent.

result = client.agents.test("agent_uid")
# {"success": True}

Returns: dict[str, Any]

inference_providers

`client.inference_providers.list()`

List inference providers.

page = client.inference_providers.list()

Parameter	Type	Default	Description
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[InferenceProvider]

`client.inference_providers.get()`

Get a single inference provider by UID.

provider = client.inference_providers.get("provider_uid")

Returns: InferenceProvider

`client.inference_providers.create()`

Create a new inference provider.

provider = client.inference_providers.create(
    name="My SageMaker",
    provider_type="sagemaker",
    config={"endpoint": "my-endpoint", "region": "us-east-1"},
)

Parameter	Type	Required	Description
`name`	`str`	Yes	Provider name.
`provider_type`	`str`	Yes	Provider type (e.g. `"sagemaker"`, `"http"`).
`config`	`dict[str, Any]`	Yes	Provider-specific configuration.
`description`	`str \| None`	No	Provider description.
`is_active`	`bool \| None`	No	Enable or disable.
`project`	`str \| None`	No	Scope to a project UID.

Returns: InferenceProvider

`client.inference_providers.update()`

Update an inference provider.

provider = client.inference_providers.update("provider_uid", is_active=False)

Parameter	Type	Required	Description
`uid`	`str`	Yes	Provider unique identifier.
`name`	`str \| None`	No	Provider name.
`description`	`str \| None`	No	Provider description.
`provider_type`	`str \| None`	No	Provider type.
`config`	`dict[str, Any] \| None`	No	Configuration object.
`is_active`	`bool \| None`	No	Enable or disable.
`project`	`str \| None`	No	Project UID scope.

Returns: InferenceProvider

`client.inference_providers.delete()`

Delete an inference provider.

client.inference_providers.delete("provider_uid")

Returns: None

`client.inference_providers.test()`

Test connectivity to an inference provider.

result = client.inference_providers.test("provider_uid")
# {"success": True, "message": "Connection successful", "tested_at": "2026-02-24T..."}

Returns: dict[str, Any]

auto_label_jobs

`client.auto_label_jobs.list()`

List auto-label jobs with optional filtering.

page = client.auto_label_jobs.list(project="proj_uid", limit=20)

Parameter	Type	Default	Description
`project`	`str \| None`	`None`	Filter by project UID.
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[AutoLabelJob]

`client.auto_label_jobs.get()`

Get a single auto-label job by UID.

job = client.auto_label_jobs.get("job_uid")
print(f"Progress: {job.progress_pct}%")

Returns: AutoLabelJob

`client.auto_label_jobs.create()`

Create a new auto-label job for a project.

job = client.auto_label_jobs.create(
    "proj_uid",
    model_type="sam3",
    confidence_threshold=0.85,
    labels=["car", "truck"],
)

Parameter	Type	Required	Description
`project_uid`	`str`	Yes	Project to auto-label.
`model_type`	`str \| None`	No	Model type to use.
`confidence_threshold`	`float \| None`	No	Minimum confidence score.
`labels`	`list[str] \| None`	No	Labels to apply.
`dry_run`	`bool \| None`	No	Preview without applying labels.

Returns: AutoLabelJob

`client.auto_label_jobs.cancel()`

Cancel a running auto-label job.

client.auto_label_jobs.cancel("job_uid")

Returns: None

quality_targets

All quality target methods are scoped to a project.

`client.quality_targets.list()`

List quality targets for a project.

page = client.quality_targets.list("proj_uid")

Parameter	Type	Default	Description
`project_uid`	`str`	—	Project unique identifier.
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[QualityTarget]

`client.quality_targets.get()`

Get a single quality target.

target = client.quality_targets.get("proj_uid", "target_uid")

Parameter	Type	Description
`project_uid`	`str`	Project unique identifier.
`uid`	`str`	Quality target unique identifier.

Returns: QualityTarget

`client.quality_targets.create()`

Create a quality target for a project.

target = client.quality_targets.create(
    "proj_uid",
    name="Accuracy Target",
    metric="accuracy",
    threshold=0.95,
    operator="gte",
    severity="critical",
    notify_webhook=True,
    notify_emails=["alerts@example.com"],
)

Parameter	Type	Required	Description
`project_uid`	`str`	Yes	Project unique identifier.
`name`	`str`	Yes	Target name.
`metric`	`str`	Yes	Metric to track.
`threshold`	`float`	Yes	Threshold value.
`operator`	`str \| None`	No	Comparison operator (`"gte"`, `"lte"`, etc.).
`severity`	`str \| None`	No	Severity level.
`is_active`	`bool \| None`	No	Enable or disable.
`notify_webhook`	`bool \| None`	No	Send webhook on breach.
`notify_emails`	`list[str] \| None`	No	Email addresses for breach alerts.

Returns: QualityTarget

`client.quality_targets.update()`

Update a quality target.

target = client.quality_targets.update("proj_uid", "target_uid", threshold=0.9)

Parameter	Type	Required	Description
`project_uid`	`str`	Yes	Project unique identifier.
`uid`	`str`	Yes	Quality target unique identifier.
`name`	`str \| None`	No	Target name.
`metric`	`str \| None`	No	Metric to track.
`threshold`	`float \| None`	No	Threshold value.
`operator`	`str \| None`	No	Comparison operator.
`severity`	`str \| None`	No	Severity level.
`is_active`	`bool \| None`	No	Enable or disable.
`notify_webhook`	`bool \| None`	No	Send webhook on breach.
`notify_emails`	`list[str] \| None`	No	Email addresses for breach alerts.

Returns: QualityTarget

`client.quality_targets.delete()`

Delete a quality target.

client.quality_targets.delete("proj_uid", "target_uid")

Returns: None

`client.quality_targets.evaluate()`

Evaluate all quality targets for a project and return current results.

results = client.quality_targets.evaluate("proj_uid")
for r in results:
    print(f"{r.name}: {r.current_value} (breached: {r.is_breached})")

Returns: list[QualityTargetEvaluation]

consensus

All consensus methods are scoped to a project.

`client.consensus.get_summary()`

Get consensus summary statistics for a project.

summary = client.consensus.get_summary("proj_uid")
print(f"Mean score: {summary.mean_score}")
print(f"Items with consensus: {summary.items_with_consensus}/{summary.total_items}")

Returns: ConsensusSummary

`client.consensus.list_scores()`

List consensus scores for individual items.

page = client.consensus.list_scores("proj_uid", limit=50)

Parameter	Type	Default	Description
`project_uid`	`str`	—	Project unique identifier.
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[ConsensusScore]

`client.consensus.compute()`

Trigger consensus computation for a project.

result = client.consensus.compute("proj_uid")
print(result.status)  # "started"

Returns: ConsensusComputeResult

`client.consensus.get_config()`

Get consensus configuration for a project.

config = client.consensus.get_config("proj_uid")
print(f"IoU threshold: {config.iou_threshold}")

Returns: ConsensusConfig

`client.consensus.update_config()`

Update consensus configuration. Uses PUT (full replacement).

config = client.consensus.update_config(
    "proj_uid",
    iou_threshold=0.7,
    min_agreement_ratio=0.8,
    min_annotations=3,
)

Parameter	Type	Required	Description
`project_uid`	`str`	Yes	Project unique identifier.
`iou_threshold`	`float \| None`	No	IoU threshold for matching.
`min_agreement_ratio`	`float \| None`	No	Minimum agreement ratio.
`min_annotations`	`int \| None`	No	Minimum annotations required.

Returns: ConsensusConfig

webhooks

`client.webhooks.list()`

List webhook subscriptions.

page = client.webhooks.list()

Parameter	Type	Default	Description
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[Webhook]

`client.webhooks.get()`

Get a single webhook by UID.

webhook = client.webhooks.get("webhook_uid")

Returns: Webhook

`client.webhooks.create()`

Create a new webhook subscription.

webhook = client.webhooks.create(
    target_url="https://example.com/webhook",
    events=["task.completed", "export.ready"],
    is_active=True,
)

Parameter	Type	Required	Description
`target_url`	`str`	Yes	URL to receive events.
`events`	`list[str]`	Yes	Events to subscribe to.
`is_active`	`bool \| None`	No	Enable the webhook (default: `True`).
`secret`	`str \| None`	No	HMAC signing secret (auto-generated if omitted).

Returns: Webhook

`client.webhooks.update()`

Update a webhook subscription.

webhook = client.webhooks.update("webhook_uid", is_active=False)

Parameter	Type	Required	Description
`uid`	`str`	Yes	Webhook unique identifier.
`target_url`	`str \| None`	No	URL to receive events.
`events`	`list[str] \| None`	No	Events to subscribe to.
`is_active`	`bool \| None`	No	Enable or disable.

Returns: Webhook

`client.webhooks.delete()`

Delete a webhook subscription.

client.webhooks.delete("webhook_uid")

Returns: None

`client.webhooks.test()`

Send a test event to a webhook.

result = client.webhooks.test("webhook_uid")
# {"success": True}

Returns: dict[str, Any]

webhook_deliveries

`client.webhook_deliveries.list()`

List webhook delivery attempts.

page = client.webhook_deliveries.list(limit=20)

Parameter	Type	Default	Description
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[WebhookDelivery]

`client.webhook_deliveries.get()`

Get a single webhook delivery by UID.

delivery = client.webhook_deliveries.get("delivery_uid")
print(f"Status: {delivery.status}, Attempts: {delivery.attempts}")

Returns: WebhookDelivery

organizations

`client.organizations.list()`

List organizations.

page = client.organizations.list(limit=20)

Parameter	Type	Default	Description
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[Organization]

`client.organizations.get()`

Get a single organization by slug.

org = client.organizations.get("my-org")

Parameter	Type	Description
`slug`	`str`	Organization slug.

Returns: Organization

`client.organizations.create()`

Create a new organization.

org = client.organizations.create(
    name="My Team",
    visibility="private",
    industry="technology",
)

Parameter	Type	Required	Description
`name`	`str`	Yes	Organization name.
`description`	`str \| None`	No	Organization description.
`logo`	`str \| None`	No	Logo URL.
`website`	`str \| None`	No	Website URL.
`industry`	`str \| None`	No	Industry (use `industry_choices()` for options).
`email`	`str \| None`	No	Contact email.
`phone`	`str \| None`	No	Contact phone.
`visibility`	`str \| None`	No	`"public"` or `"private"`.

Returns: Organization

`client.organizations.update()`

Update an organization.

org = client.organizations.update("my-org", name="New Name", description="Updated desc")

Parameter	Type	Required	Description
`slug`	`str`	Yes	Organization slug.
`name`	`str \| None`	No	Organization name.
`description`	`str \| None`	No	Description.
`logo`	`str \| None`	No	Logo URL.
`website`	`str \| None`	No	Website URL.
`handle`	`str \| None`	No	Organization handle.
`industry`	`str \| None`	No	Industry.
`email`	`str \| None`	No	Contact email.
`phone`	`str \| None`	No	Contact phone.
`visibility`	`str \| None`	No	Visibility setting.

Returns: Organization

`client.organizations.delete()`

Deleting an organization is irreversible and may cascade-delete all associated projects, datasets, and other resources. Double-check the organization slug before calling this method.

Delete an organization.

client.organizations.delete("my-org")

Returns: None

`client.organizations.industry_choices()`

Get the list of valid industry choices.

choices = client.organizations.industry_choices()

Returns: Dict[str, Any] — A dictionary containing industry choices.

`client.organizations.list_members()`

List members of an organization.

page = client.organizations.list_members("my-org", role="admin")

Parameter	Type	Default	Description
`slug`	`str`	—	Organization slug.
`search`	`str \| None`	`None`	Search by name or email.
`role`	`str \| None`	`None`	Filter by role.
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[OrganizationMember]

`client.organizations.remove_member()`

Remove a member from an organization.

client.organizations.remove_member("my-org", "user_uid")

Returns: None

`client.organizations.update_member_role()`

Update a member’s role in an organization.

client.organizations.update_member_role("my-org", "user_uid", role="admin")

Returns: None

`client.organizations.leave()`

Leave an organization (as the current user).

client.organizations.leave("my-org")

Returns: None

`client.organizations.transfer_ownership()`

Transfer organization ownership to another member.

client.organizations.transfer_ownership("my-org", new_owner_uid="new_owner_uid")

Returns: None

`client.organizations.list_invitations()`

List pending invitations for an organization.

invitations = client.organizations.list_invitations("my-org")

Returns: list[Invitation]

`client.organizations.create_invitation()`

Invite a user to an organization.

invitation = client.organizations.create_invitation("my-org", email="user@example.com", role="member")

Parameter	Type	Required	Description
`slug`	`str`	Yes	Organization slug.
`email`	`str`	Yes	Email to invite.
`role`	`str \| None`	No	Role for the invitee.

Returns: Invitation

`client.organizations.resend_invitation()`

Resend a pending invitation.

client.organizations.resend_invitation("my-org", "invitation_uid")

Returns: None

`client.organizations.cancel_invitation()`

Cancel a pending invitation.

client.organizations.cancel_invitation("my-org", "invitation_uid")

Returns: None

`client.organizations.list_teams()`

List teams in an organization.

teams = client.organizations.list_teams("my-org")

Returns: list[Team]

`client.organizations.create_team()`

Create a team in an organization.

team = client.organizations.create_team("my-org", name="ML Engineers", description="Machine learning team")

Parameter	Type	Required	Description
`slug`	`str`	Yes	Organization slug.
`name`	`str`	Yes	Team name.
`description`	`str \| None`	No	Team description.
`color`	`str \| None`	No	Team color.

Returns: Team

`client.organizations.get_team()`

Get a team by slug.

team = client.organizations.get_team("my-org", "team_slug")

Returns: Team

`client.organizations.update_team()`

Update a team.

team = client.organizations.update_team("my-org", "team_slug", name="Updated Name")

Returns: Team

`client.organizations.delete_team()`

Delete a team.

client.organizations.delete_team("my-org", "team_slug")

Returns: None

`client.organizations.list_team_members()`

List members of a team.

members = client.organizations.list_team_members("my-org", "team_slug")

Returns: list[TeamMember]

`client.organizations.add_team_member()`

Add a member to a team.

member = client.organizations.add_team_member("my-org", "team_slug", user_uid="user_uid", role="member")

Returns: TeamMember

`client.organizations.remove_team_member()`

Remove a member from a team.

client.organizations.remove_team_member("my-org", "team_slug", "user_uid")

Returns: None

`client.organizations.update_team_member_role()`

Update a team member’s role.

client.organizations.update_team_member_role("my-org", "team_slug", "user_uid", role="lead")

Returns: None

slices

`client.slices.list()`

List slices for an owner.

page = client.slices.list("my-org", limit=20)

Parameter	Type	Default	Description
`owner`	`str`	—	Owner slug (organization or user).
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[Slice]

`client.slices.get()`

Get a single slice.

s = client.slices.get("my-org", "my-slice")

Parameter	Type	Description
`owner`	`str`	Owner slug.
`slug`	`str`	Slice slug.

Returns: Slice

`client.slices.create()`

Create a new slice.

s = client.slices.create(
    name="Validation Set",
    visibility="private",
    sub_slices=[{"dataset": "dataset_uid", "filters": {}}],
    organization="my-org",
)

Parameter	Type	Required	Description
`name`	`str`	Yes	Slice name.
`visibility`	`str`	Yes	`"public"` or `"private"`.
`sub_slices`	`list[dict[str, Any]]`	Yes	Sub-slice definitions (dataset + filters).
`organization`	`str \| None`	No	Organization slug to own the slice.
`random_selection_count`	`int \| None`	No	Randomly select N items.

Returns: Slice

`client.slices.list_items()`

List items in a slice.

page = client.slices.list_items("my-org", "my-slice", limit=50)

Parameter	Type	Default	Description
`owner`	`str`	—	Owner slug.
`slug`	`str`	—	Slice slug.
`limit`	`int \| None`	`None`	Maximum results per page.
`cursor`	`str \| None`	`None`	Pagination cursor.

Returns: CursorPage[SliceItem]

`client.slices.get_item()`

Get a single item within a slice.

item = client.slices.get_item("my-org", "my-slice", "item_uid")

Parameter	Type	Description
`owner`	`str`	Owner slug.
`slug`	`str`	Slice slug.
`item_uid`	`str`	Item unique identifier.

Returns: SliceItem

Types

Dataset

class Dataset(BaseModel):
    uid: str
    name: str
    slug: str
    item_count: int = 0
    data_type: str | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

DatasetItem

class DatasetItem(BaseModel):
    id: int | None = None
    uid: str
    key: str | None = None
    dataset: str | None = None
    url: str | None = None
    gpu_texture_url: str | None = None
    thumbnails: list[str] | None = None
    video_thumbnail: str | None = None
    metadata: dict[str, Any] | None = None
    export_snippet: dict[str, Any] | None = None
    annotations: dict[str, Any] | None = None
    crop_data: dict[str, Any] | None = None
    related_items: list[dict[str, Any]] | None = None
    related_sequence_uid: str | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

DatasetSequence

class DatasetSequence(BaseModel):
    uid: str
    key: str | None = None
    custom_uuid: str | None = None
    status: str | None = None
    featured_image: str | None = None
    number_of_frames: int | None = None
    views: list[dict[str, Any]] | None = None
    crop_data: dict[str, Any] | None = None
    predefined_labels: list[dict[str, Any]] | None = None
    frames: list[dict[str, Any]] | None = None
    metrics: dict[str, Any] | None = None
    dataset_uid: str | None = None
    allow_lidar_calibration: bool | None = None
    lidar_calibration_enabled: bool | None = None
    camera_calibration_enabled: bool | None = None

Project

class Project(BaseModel):
    uid: str
    name: str
    status: str | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

Task

class Task(BaseModel):
    uid: str
    type: str | None = None
    name: str | None = None
    status: str | None = None
    project: str | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

Export

class Export(BaseModel):
    uid: str
    status: str | None = None
    download_url: str | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

StorageConfig

class StorageConfig(BaseModel):
    uid: str
    name: str
    provider: str
    s3_bucket_name: str | None = None
    s3_bucket_region: str | None = None
    s3_bucket_prefix: str | None = None
    s3_is_accelerated: bool = False
    gc_storage_bucket_name: str | None = None
    gc_storage_prefix: str | None = None
    is_verified: bool = False
    last_verified_at: datetime | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

Agent

class Agent(BaseModel):
    uid: str
    name: str
    description: str | None = None
    events: list[str] = []
    callback_url: str | None = None
    is_active: bool = True
    project: str | None = None
    task_types: list[str] = []
    secret: str | None = None
    execution_stats: dict[str, int] | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

AgentExecution

class AgentExecution(BaseModel):
    uid: str
    registration: str
    event_type: str
    task: str | None = None
    result: str | None = None
    status: str | None = None
    action: str | None = None
    event_payload: dict[str, Any] | None = None
    response_payload: dict[str, Any] | None = None
    error_message: str | None = None
    started_at: datetime | None = None
    completed_at: datetime | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

InferenceProvider

class InferenceProvider(BaseModel):
    uid: str
    name: str
    description: str | None = None
    provider_type: str | None = None
    config: dict[str, Any] | None = None
    is_active: bool = True
    project: str | None = None
    last_test_at: datetime | None = None
    last_test_success: bool | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

AutoLabelJob

class AutoLabelJob(BaseModel):
    uid: str
    status: str | None = None
    model_type: str | None = None
    confidence_threshold: float | None = None
    labels: list[str] = []
    dry_run: bool = False
    total_items: int = 0
    processed_items: int = 0
    successful_items: int = 0
    failed_items: int = 0
    skipped_items: int = 0
    progress_pct: float | None = None
    error_message: str | None = None
    summary: dict[str, Any] | None = None
    started_at: datetime | None = None
    completed_at: datetime | None = None
    created_at: datetime | None = None

QualityTarget

class QualityTarget(BaseModel):
    uid: str
    name: str
    metric: str
    operator: str
    threshold: float
    severity: str | None = None
    is_active: bool = True
    notify_webhook: bool = True
    notify_emails: list[str] = []
    last_evaluated_at: datetime | None = None
    last_value: float | None = None
    is_breached: bool = False
    breach_count: int = 0
    last_breached_at: datetime | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

QualityTargetEvaluation

class QualityTargetEvaluation(BaseModel):
    uid: str
    name: str
    metric: str
    threshold: float
    operator: str
    current_value: float
    is_breached: bool
    severity: str | None = None

ConsensusConfig

class ConsensusConfig(BaseModel):
    uid: str
    iou_threshold: float = 0.5
    min_agreement_ratio: float = 0.5
    min_annotations: int = 2
    created_at: datetime | None = None
    updated_at: datetime | None = None

ConsensusScore

class ConsensusScore(BaseModel):
    uid: str
    dataset_item_uid: str
    task_name: str
    score_type: str | None = None
    score: float
    annotator_count: int
    details: dict[str, Any] | None = None
    created_at: datetime | None = None

ConsensusSummary

class ConsensusSummary(BaseModel):
    mean_score: float
    median_score: float
    min_score: float
    max_score: float
    total_items: int
    items_with_consensus: int
    score_distribution: dict[str, Any] | None = None
    by_task_name: list[Any] | None = None

ConsensusComputeResult

class ConsensusComputeResult(BaseModel):
    status: str
    message: str

Webhook

class Webhook(BaseModel):
    uid: str
    target_url: str
    events: list[str] = []
    is_active: bool = True
    secret: str | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

WebhookDelivery

class WebhookDelivery(BaseModel):
    uid: str
    subscription: str
    event_type: str
    payload: dict[str, Any] | None = None
    response_status: int | None = None
    response_body: str | None = None
    attempts: int = 0
    next_retry_at: datetime | None = None
    status: str | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

Organization

class Organization(BaseModel):
    uid: str
    name: str
    slug: str
    handle: str | None = None
    description: str | None = None
    logo: str | None = None
    website: str | None = None
    industry: str | None = None
    email: str | None = None
    phone: str | None = None
    visibility: str | None = None
    plan: str | None = None
    is_verified: bool = False
    is_active: bool = True
    member_count: int | None = None
    team_count: int | None = None
    dataset_count: int | None = None
    project_count: int | None = None
    slice_count: int | None = None
    role: str | None = None
    joined_at: datetime | None = None
    allowed_domains: list[str] | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

OrganizationMember

class OrganizationMember(BaseModel):
    user_uid: str
    username: str | None = None
    email: str | None = None
    full_name: str | None = None
    picture: str | None = None
    role: str | None = None
    created_at: datetime | None = None

Invitation

class Invitation(BaseModel):
    uid: str
    organization_name: str | None = None
    organization_slug: str | None = None
    invited_email: str
    invited_by_username: str | None = None
    role: str | None = None
    status: str | None = None
    expires_at: datetime | None = None
    is_expired: bool | None = None
    accept_url: str | None = None
    copy_link: str | None = None
    created_at: datetime | None = None

Team

class Team(BaseModel):
    uid: str
    name: str
    slug: str | None = None
    description: str | None = None
    color: str | None = None
    organization_uid: str | None = None
    organization_name: str | None = None
    organization_slug: str | None = None
    member_count: int | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

TeamMember

class TeamMember(BaseModel):
    user_uid: str
    username: str | None = None
    email: str | None = None
    full_name: str | None = None
    picture: str | None = None
    role: str | None = None
    created_at: datetime | None = None

Slice

class Slice(BaseModel):
    uid: str
    name: str
    slug: str | None = None
    owner_name: str | None = None
    organization: dict[str, Any] | None = None
    visibility: str | None = None
    status: str | None = None
    item_count: int | None = None
    sub_slices: list[dict[str, Any]] | None = None
    source_data: Any | None = None
    featured_slice_item_urls: list[str] | None = None

SliceItem

class SliceItem(BaseModel):
    id: int | None = None
    uid: str
    key: str | None = None
    dataset: str | None = None
    url: str | None = None
    gpu_texture_url: str | None = None
    thumbnails: list[str] | None = None
    video_thumbnail: str | None = None
    metadata: dict[str, Any] | None = None
    export_snippet: dict[str, Any] | None = None
    annotations: dict[str, Any] | None = None
    crop_data: dict[str, Any] | None = None
    related_items: list[dict[str, Any]] | None = None
    related_sequence_uid: str | None = None
    created_at: datetime | None = None
    updated_at: datetime | None = None

Pagination

CursorPage

All list() methods return a CursorPage[T]:

@dataclass
class CursorPage(Generic[T]):
    items: list[T]
    next_cursor: str | None = None
    previous_cursor: str | None = None

Property	Type	Description
`items`	`list[T]`	Items on the current page.
`next_cursor`	`str \| None`	Cursor for the next page. `None` if last page.
`previous_cursor`	`str \| None`	Cursor for the previous page.
`has_more`	`bool`	`True` if `next_cursor` is not `None`.

Supports iteration and len():

page = client.datasets.list(limit=20)
for dataset in page:
    print(dataset.name)

print(f"{len(page)} datasets on this page")

if page.has_more:
    next_page = client.datasets.list(limit=20, cursor=page.next_cursor)

Errors

All errors inherit from AvalaError:

from avala.errors import (
    AvalaError,
    AuthenticationError,
    NotFoundError,
    RateLimitError,
    ValidationError,
    ServerError,
)

Error	HTTP Status	Attributes
`AvalaError`	Any	`message`, `status_code`, `body`
`AuthenticationError`	401	Inherits from `AvalaError`
`NotFoundError`	404	Inherits from `AvalaError`
`RateLimitError`	429	`retry_after: float \| None`
`ValidationError`	400 / 422	`details: list[Any]`
`ServerError`	5xx	Inherits from `AvalaError`

try:
    dataset = client.datasets.get("nonexistent")
except NotFoundError as e:
    print(f"Not found: {e.message}")
except RateLimitError as e:
    print(f"Retry after {e.retry_after}s")
except AvalaError as e:
    print(f"Error ({e.status_code}): {e.message}")

Environment Variables

Variable	Description	Default
`AVALA_API_KEY`	API key (used when `api_key` not passed to client).	Required
`AVALA_BASE_URL`	Base API URL override.	`https://api.avala.ai/api/v1`

SDKs

SDK Reference

​Client

​avala.Client

​avala.AsyncClient

​Resources

​datasets

​client.datasets.create()

​client.datasets.list()

​client.datasets.get()

​client.datasets.list_items()

​client.datasets.get_item()

​client.datasets.list_sequences()

​client.datasets.get_sequence()

​projects

​client.projects.list()

​client.projects.get()

​tasks

​client.tasks.list()

​client.tasks.get()

​exports

​client.exports.list()

​client.exports.create()

​client.exports.get()

​storage_configs

​client.storage_configs.list()

​client.storage_configs.create()

​client.storage_configs.test()

​client.storage_configs.delete()

​agents

​client.agents.list()

​client.agents.get()

​client.agents.create()

​client.agents.update()

​client.agents.delete()

​client.agents.list_executions()

​client.agents.test()

​inference_providers

​client.inference_providers.list()

​client.inference_providers.get()

​client.inference_providers.create()

​client.inference_providers.update()

​client.inference_providers.delete()

​client.inference_providers.test()

​auto_label_jobs

​client.auto_label_jobs.list()

​client.auto_label_jobs.get()

​client.auto_label_jobs.create()

​client.auto_label_jobs.cancel()

​quality_targets

​client.quality_targets.list()

​client.quality_targets.get()

​client.quality_targets.create()

​client.quality_targets.update()

​client.quality_targets.delete()

​client.quality_targets.evaluate()

​consensus

​client.consensus.get_summary()

​client.consensus.list_scores()

​client.consensus.compute()

​client.consensus.get_config()

​client.consensus.update_config()

​webhooks

​client.webhooks.list()

​client.webhooks.get()

​client.webhooks.create()

​client.webhooks.update()

​client.webhooks.delete()

​client.webhooks.test()

​webhook_deliveries

​client.webhook_deliveries.list()

​client.webhook_deliveries.get()

​organizations

​client.organizations.list()

​client.organizations.get()

​client.organizations.create()

​client.organizations.update()

​client.organizations.delete()

​client.organizations.industry_choices()

​client.organizations.list_members()

Client

`avala.Client`

`avala.AsyncClient`

Resources

datasets

`client.datasets.create()`

`client.datasets.list()`

`client.datasets.get()`

`client.datasets.list_items()`

`client.datasets.get_item()`

`client.datasets.list_sequences()`

`client.datasets.get_sequence()`

projects

`client.projects.list()`

`client.projects.get()`

tasks

`client.tasks.list()`

`client.tasks.get()`

exports

`client.exports.list()`

`client.exports.create()`

`client.exports.get()`

storage_configs

`client.storage_configs.list()`

`client.storage_configs.create()`

`client.storage_configs.test()`

`client.storage_configs.delete()`

agents

`client.agents.list()`

`client.agents.get()`

`client.agents.create()`

`client.agents.update()`

`client.agents.delete()`

`client.agents.list_executions()`

`client.agents.test()`

inference_providers

`client.inference_providers.list()`

`client.inference_providers.get()`

`client.inference_providers.create()`

`client.inference_providers.update()`

`client.inference_providers.delete()`

`client.inference_providers.test()`

auto_label_jobs

`client.auto_label_jobs.list()`

`client.auto_label_jobs.get()`

`client.auto_label_jobs.create()`

`client.auto_label_jobs.cancel()`

quality_targets

`client.quality_targets.list()`

`client.quality_targets.get()`

`client.quality_targets.create()`

`client.quality_targets.update()`

`client.quality_targets.delete()`

`client.quality_targets.evaluate()`

consensus

`client.consensus.get_summary()`

`client.consensus.list_scores()`

`client.consensus.compute()`

`client.consensus.get_config()`

`client.consensus.update_config()`

webhooks

`client.webhooks.list()`

`client.webhooks.get()`

`client.webhooks.create()`

`client.webhooks.update()`

`client.webhooks.delete()`

`client.webhooks.test()`

webhook_deliveries

`client.webhook_deliveries.list()`

`client.webhook_deliveries.get()`

organizations

`client.organizations.list()`

`client.organizations.get()`

`client.organizations.create()`

`client.organizations.update()`

`client.organizations.delete()`

`client.organizations.industry_choices()`

`client.organizations.list_members()`

`client.organizations.remove_member()`

`client.organizations.update_member_role()`