Skip to main content
The Avala MCP server provides tools for managing all aspects of your workspace. Tools are organized by category and availability depends on whether mutations are enabled.
Tools marked with Write require AVALA_MCP_ENABLE_MUTATIONS=true. All other tools are available in read-only mode.

​
Datasets

Manage annotation datasets and their configurations.

​
list_datasets

List all datasets in your workspace with filtering options. Parameters:
  • dataType (optional) - Filter by type: image, video, lidar, mcap, or splat
  • name (optional) - Filter by name (substring match, case-insensitive)
  • status (optional) - Filter by status: creating or created
  • visibility (optional) - Filter by visibility: private or public
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
get_dataset

Get detailed information about a specific dataset. Parameters:
  • uid (required) - Dataset UUID

​
create_dataset Write

Create a new dataset for annotation. Parameters:
  • name (required) - Display name for the dataset
  • slug (required) - URL-friendly identifier
  • dataType (required) - Type: image, video, lidar, or mcap
  • isSequence (optional) - Whether dataset contains sequences (default: false)
  • visibility (optional) - private or public (default: private)
  • createMetadata (optional) - Whether to create metadata (default: true)
  • providerConfig (optional) - Cloud storage configuration
  • ownerName (optional) - Owner username or email

​
Projects

Access annotation project details and status.

​
list_projects

List all annotation projects with status and progress. Parameters:
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
get_project

Get full project details including configuration. Parameters:
  • uid (required) - Project UUID

​
Tasks

Query and filter annotation tasks.

​
list_tasks

List tasks with optional filtering. Parameters:
  • project (optional) - Filter by project UID
  • status (optional) - Filter by task status
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
get_task

Get detailed information about a specific task. Parameters:
  • uid (required) - Task UUID

​
Statistics

Access workspace-wide statistics and metrics.

​
get_workspace_stats

Get summary of workspace usage including dataset, project, and export counts. Parameters: None

​
Agents

Manage automation agents for event-driven workflows.

​
list_agents

List all automation agents in your workspace. Parameters:
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
get_agent

Get detailed information about a specific agent. Parameters:
  • uid (required) - Agent UUID

​
create_agent Write

Create a new automation agent. Parameters:
  • name (required) - Name of the agent
  • events (required) - Array of event types to subscribe to
  • callbackUrl (optional) - URL to receive event callbacks
  • description (optional) - Description of the agent
  • project (optional) - Project UID to scope the agent
  • taskTypes (optional) - Array of task types the agent handles

​
delete_agent Write

Delete an automation agent. Parameters:
  • uid (required) - Agent UUID to delete

​
Annotation issues

Manage quality control issues and QC tools.

​
list_annotation_issues_by_sequence

List annotation issues for a specific sequence. Parameters:
  • sequenceUid (required) - Sequence UID
  • datasetItemUid (optional) - Filter by dataset item
  • projectUid (optional) - Filter by project

​
list_annotation_issues_by_dataset

List annotation issues for a dataset. Parameters:
  • owner (required) - Dataset owner
  • datasetSlug (required) - Dataset slug
  • sequenceUid (optional) - Filter by sequence

​
get_annotation_issue_metrics

Get annotation issue metrics for a dataset. Parameters:
  • owner (required) - Dataset owner
  • datasetSlug (required) - Dataset slug
  • sequenceUid (optional) - Filter by sequence

​
list_qc_tools

List available QC annotation tools for a dataset type. Parameters:
  • datasetType (required) - Dataset type: image, video, or lidar

​
create_annotation_issue Write

Create a new annotation issue on a sequence. Parameters:
  • sequenceUid (required) - Sequence UID
  • toolUid (required) - Annotation tool UID
  • problemUid (required) - Issue problem UID
  • datasetItemUid (optional) - Dataset item UID
  • projectUid (optional) - Project UID
  • priority (optional) - lowest, low, medium, high, or highest
  • severity (optional) - critical or moderate
  • description (optional) - Issue description
  • wrongClass (optional) - Incorrect class label
  • correctClass (optional) - Correct class label
  • shouldReAnnotate (optional) - Whether to re-annotate
  • shouldDelete (optional) - Whether to delete annotation
  • framesAffected (optional) - Frames affected by issue
  • coordinates (optional) - Issue coordinates
  • queryParams (optional) - Additional query parameters
  • objectUid (optional) - Annotation object UID

​
update_annotation_issue Write

Update an existing annotation issue. Parameters:
  • sequenceUid (required) - Sequence UID
  • issueUid (required) - Issue UID
  • status (optional) - open, relabeling, in_review, completed, cant_reproduce, awaiting_feedback, or no_action_taken
  • priority (optional) - Issue priority
  • severity (optional) - Issue severity
  • description (optional) - Issue description
  • toolUid (optional) - Annotation tool UID
  • problemUid (optional) - Issue problem UID
  • wrongClass (optional) - Incorrect class
  • framesAffected (optional) - Affected frames

​
delete_annotation_issue Write

Delete an annotation issue. Parameters:
  • sequenceUid (required) - Sequence UID
  • issueUid (required) - Issue UID to delete

​
Quality targets

Manage quality targets and evaluation.

​
list_quality_targets

List quality targets configured for a project. Parameters:
  • projectUid (required) - Project UUID
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
evaluate_quality Write

Evaluate all quality targets for a project. Parameters:
  • projectUid (required) - Project UUID to evaluate

​
Consensus

Access consensus scores and trigger computation.

​
get_consensus_summary

Get consensus summary including mean/median scores and distribution. Parameters:
  • projectUid (required) - Project UUID

​
compute_consensus Write

Trigger consensus computation for a project. Parameters:
  • projectUid (required) - Project UUID

​
Webhooks

Manage webhook subscriptions for event notifications.

​
list_webhooks

List all webhook subscriptions. Parameters:
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
create_webhook Write

Create a new webhook subscription. Parameters:
  • targetUrl (required) - URL to receive deliveries
  • events (required) - Array of event types to subscribe to

​
delete_webhook Write

Delete a webhook subscription. Parameters:
  • uid (required) - Webhook UUID to delete

​
Storage configurations

Manage cloud storage integrations.

​
list_storage_configs

List all storage configurations. Parameters:
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
test_storage_config

Test connectivity for a storage configuration. Parameters:
  • uid (required) - Storage config UUID to test

​
create_storage_config Write

Create a new storage configuration (S3 or GCS). Parameters:
  • name (required) - Configuration name
  • provider (required) - Provider type: s3 or gcs
  • s3BucketName (optional) - S3 bucket name
  • s3BucketRegion (optional) - S3 region
  • s3BucketPrefix (optional) - S3 key prefix
  • s3AccessKeyId (optional) - S3 access key ID
  • s3SecretAccessKey (optional) - S3 secret key
  • s3IsAccelerated (optional) - Enable S3 Transfer Acceleration
  • gcStorageBucketName (optional) - GCS bucket name
  • gcStoragePrefix (optional) - GCS prefix
  • gcStorageAuthJsonContent (optional) - GCS service account JSON
Credentials are transmitted securely and stored server-side. The MCP server does not store credentials locally.

​
delete_storage_config Write

Delete a storage configuration. Parameters:
  • uid (required) - Storage config UUID to delete

​
Exports

Trigger and monitor data exports.

​
list_exports

List all exports with formats and download URLs. Parameters:
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
get_export_status

Check export processing status. Parameters:
  • uid (required) - Export UUID

​
create_export Write

Trigger a new export for a dataset or project. Parameters:
  • project (optional) - Project UID to export
  • dataset (optional) - Dataset UID to export
Provide either project or dataset, not both.

​
Organizations

Access organization information and membership.

​
list_organizations

List all organizations you are a member of. Parameters:
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
get_organization

Get detailed organization information. Parameters:
  • slug (required) - Organization slug identifier

​
Slices

Access dataset slices and filters.

​
list_slices

List slices for an owner (user or organization). Parameters:
  • owner (required) - Owner name or organization slug
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
get_slice

Get detailed information about a slice. Parameters:
  • owner (required) - Owner name or organization slug
  • slug (required) - Slice slug

​
Fleet management

Monitor and manage IoT devices and data collection.

​
fleet_list_devices

List fleet devices with filtering. Parameters:
  • status (optional) - Filter by status: online, offline, or maintenance
  • type (optional) - Filter by device type
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
fleet_get_device

Get detailed device information. Parameters:
  • uid (required) - Device UUID

​
fleet_list_recordings

List fleet recordings with filtering. Parameters:
  • device (optional) - Filter by device UID
  • status (optional) - Filter by recording status
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
fleet_get_recording

Get detailed recording information. Parameters:
  • uid (required) - Recording UUID

​
fleet_list_events

List fleet events with filtering. Parameters:
  • recording (optional) - Filter by recording UID
  • device (optional) - Filter by device UID
  • type (optional) - Filter by event type
  • severity (optional) - Filter by severity
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
fleet_list_alerts

List fleet alerts with filtering. Parameters:
  • status (optional) - Filter by status: open, acknowledged, or resolved
  • severity (optional) - Filter by severity: info, warning, error, or critical
  • device (optional) - Filter by device UID
  • rule (optional) - Filter by rule UID
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
fleet_list_rules

List fleet monitoring rules. Parameters:
  • enabled (optional) - Filter by enabled status
  • limit (optional) - Maximum number of results
  • cursor (optional) - Pagination cursor

​
fleet_register_device Write

Register a new fleet device. Parameters:
  • name (required) - Device name
  • type (required) - Device type
  • firmwareVersion (optional) - Firmware version
  • tags (optional) - Array of tags

​
fleet_acknowledge_alert Write

Acknowledge a fleet alert. Parameters:
  • uid (required) - Alert UUID to acknowledge

​
Response format

All tools return JSON responses. Successful operations include the requested data or confirmation: 
{
  "content": [
    {
      "type": "text",
      "text": "{\"uid\":\"abc123\",\"name\":\"My Dataset\",...}"
    }
  ]
}
Paginated responses include cursor information for fetching additional pages: 
{
  "items": [...],
  "cursor": "eyJpZCI6MTIzfQ==",
  "hasMore": true
}

Build docs developers (and LLMs) love

Get started for freeTalk to us