proto/clarifai/api/service.proto

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Identify annotation filter by id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

List annotations for these IDs

List annotations for these Input IDs
Note that ‘ids’ are optional but if the are provided the number and order in
‘ids’ and ‘input_ids’ should match
If you do not specify ‘ids’ all the annotations for ‘input_ids’ are returned
If a you do not specify both ‘input_ids’ and ‘ids’ all the annotations in the app are returned

Only return the annotations that has one of these user IDs, effectively operating as an
OR among them to filter down the results.
If model_version_ids are also provided these user_ids are OR’d with them as well since
annotations are either provided by users or model versions and we want the union of any
provided user or model version annotations in the results of ListAnnotations request.
If no user_ids are provided then annotations from all users are returned.

Only return the annotations that has one of these model version IDs, effectively operating as an
OR among them to filter down the results.
If user_ids are also provided these model_versions_ids are OR’d with them as well since
annotations are either provided by users or model versions and we want the union of any
provided user or model version annotations in the results of ListAnnotations request.
If no model_version_ids are provided then annotations from all model versions are returned.

Set this flag to list both trusted and not trusted annotations
by default it’s listing only trusted annotations

Set this flag to return the model output for model annotations in the response.

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Flag to filter annotations by task_id

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Unique custom facing id that identifies the eval to get

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

(optional URL parameter) List of additional fields to be included in the response.
Currently supported: all, names

(optional URL parameter) Only list workers that have created trusted annotations.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) List of additional fields to be included in the response. Currently supported: all, stars

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

(optional URL parameter) List of additional fields to be included in the response. Currently supported: all, stars

Sorting options:
Whether to sort in ascending order. If false, will order in descending order.

Whether to order by the name

Whether to order by the modified_at time.
If none of the sort options is set to true, will sort by modified_at.

Whether to order by the created_at time.

Whether to order by the number of users stared the app

Filtering options:
If true, we only return apps that are handpicked by clarifai staff

If true, we only return apps that are starred by the requesting user

If true, we only return apps that are marked as a template by the app owner.

Searching options:
Specify a search parameter in order to perform keyword search on the
following fields of the application:

• id
• name
• description
• notes
• user_id (unless user_app_id.user_id is already set)

Keywords are both normalized for search (so searching for “satisfy” matches “satisfied”)
and used for partial prefix-matching (so searching for “clari” matches “clarifai”).

NOTE: Both the list of fields searched and the exact keyword matching
rules are subject to change and not guaranteed to be backwards-compatible.

Query various text fields (id, name, description, and notes) that can contain the words in the query string
Deprecated: use search instead.

Filter by the id, name and notes of the app. This supports wilcard queries like “gen*” to match “general” as an example.
Deprecated: use search instead.

Filter by the user-unique-id of the app. This supports wilcard queries like “gen*” to match “general” as an example.
Deprecated: use search instead.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

The bulk operation id

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Fuzzy match on concept ID

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

The concept’s id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

This is the language identifier.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Sorting options:
Whether to sort in ascending order. If false, will order in descending order.

Whether to order by the created_at time.

Whether to order by the number of users stared the app

If neither sort option is set to true, will sort by modified_at.

Whether to order by the external id

Filtering options:

Filter datasets by bookmark. If set, only return bookmarked datasets. Otherwise none bookmarked datasets only.

Searching options:
Specify a search parameter in order to perform keyword search on the
following fields of the dataset:

• id
• description
• notes
• user_id (unless user_app_id.user_id is already set)

Keywords are both normalized for search (so searching for “satisfy” matches “satisfied”)
and used for partial prefix-matching (so searching for “clari” matches “clarifai”).

NOTE: Both the list of fields searched and the exact keyword matching
rules are subject to change and not guaranteed to be backwards-compatible.

Fuzzy filter on dataset ID
Deprecated: use search instead.

Identify dataset by id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Identify dataset by id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Identify dataset by id.

Identify dataset input by id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Identify dataset by id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Identify dataset by id.

Identify dataset version by id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Identify dataset by id.

Identify dataset version by id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Filter by parent path. Multiple parent paths are OR-ed. Empty list results in no filter.

Filter by parent type. Multiple types are OR-ed. Empty list results in no filter.

• INPUT_TYPE: Group data examples by input type.
  Examples: images, videos, text, audio.
• CONCEPT_ID: Group data examples by concept ID.
  Examples: inputs with cat concept, inputs with dog concept.
• CONCEPTS_COUNT: Group data examples by concepts count.
  Examples: inputs with 20 concepts, inputs with 21 concepts.
• BOUNDING_BOXES_COUNT: Group data examples by bounding boxes count.
  Examples: inputs with 20 bounding boxes, inputs with 21 bounding boxes.
• POLYGONS_COUNT: Group data examples by polygons count.
  Examples: inputs with 20 polygons, inputs with 21 polygons.
• POINTS_COUNT: Group data examples by points count.
  Examples: inputs with 20 points, inputs with 21 points.
• MASKS_COUNT: Group data examples by masks count.
  Examples: inputs with 20 masks, inputs with 21 masks.
• PIXELS_COUNT: Group data examples by pixels count.
  In order to reduce the number of groups, we use bins.
  Examples for bin size = 400: inputs with [200000, 200400) pixels, inputs with [200400, 200800) pixels.
• ASPECT_RATIO: Group data examples by aspect ratio.
  In order to reduce the number of groups, we use bins.
  Examples for bin size = 0.1: inputs with [0.5, 0.6) aspect ratio, inputs with [0.6, 0.7) aspect ratio.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

If true, sort in ascending order, otherwise sort in descending order.

Whether to order by application id

Whether to order by eval metric summary.macro_avg_roc_auc

Whether to order by eval metric summary.f1

Whether to order by when the metric was created

Whether to order by eval metric summary.mean_avg_precision_iou_50

Whether to order by eval metric summary.macro_avg_precision

Whether to order by eval metric summary.macro_avg_recall

Filter on model type id

Filter on dataset ID of the dataset version specified in the metric version

Filter on dataset ID of the dataset version specified by the model version

Filter on concept IDs specified in the modele version’s output_info

Whether to show failed metrics, defaults to false

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Status code from internal codes.

• ZERO: to be revised and greatly expanded
• SUCCESS: Generic
• MOVED: Resource moved. Respond with Http status 307 and add new Location header to response
• TEAPOT: SUCCESS_WARNING_API_DEPRECATED = 10001;
  SUCCESS_WARNING_CLIENT_DEPRECATED = 10002;
• CONN_ACCOUNT_ISSUES: Clarifai Connection Codes: 11xxx

expired cc, still in trial, feature not supported in your tier

• CONN_TOKEN_INVALID: invalid auth token used. Deprecated: we should return CONN_KEY_INVALID instead now in all cases.
• CONN_CREDENTIALS_INVALID: invalid auth credentials
• CONN_EXCEED_HOURLY_LIMIT: throttle hourly limit exceeded
• CONN_EXCEED_MONTHLY_LIMIT: throttle monthly limit exceeded
• CONN_THROTTLED: throttler and billing stuff
• CONN_EXCEEDS_LIMITS: throttler and billing stuff
• CONN_INSUFFICIENT_SCOPES: api key has insufficient permissions
• CONN_KEY_INVALID: api key is invalid
• CONN_KEY_NOT_FOUND: api key not found
• CONN_BAD_REQUEST_FORMAT: multipart form parsing, broken json, etc
• CONN_DOES_NOT_EXIST: when path is bad
• CONN_INVALID_REQUEST: something wrong with a header
• CONN_METHOD_NOT_ALLOWED: when a request method is not allowed
• CONN_NO_GDPR_CONSENT: lack GDPR consent
• CONN_AUTH_METHOD_DISABLED: authentication method is disabled
• MODEL_TRAINED: Model/Custom Training related 20xxx

Custom model has been already trained.

• MODEL_TRAINING: Custom model is currently training.
• MODEL_UNTRAINED: Custom model has not yet been trained.
• MODEL_QUEUED_FOR_TRAINING: Custom model is currently in queue for training, waiting on assets to process first.
• MODEL_TRAINING_FAILED: generic err msg for any type of model training err.
• MODEL_TRAINING_NO_DATA: Custom model training had no data. FIXME(yang): deprecate this. Use the 21106 + errStatusMsg
• MODEL_TRAINING_NO_POSITIVES: Custom model training had no positive examples. FIXME(yang): deprecate this. Use the 21106 + errStatusMsg
• MODEL_TRAINING_ONE_VS_N_SINGLE_CLASS: Custom model training was ONE_VS_N but with a single class. FIXME(yang): deprecate this. Use the 21106 + errStatusMsg
• MODEL_TRAINING_TIMED_OUT: Training took longer than hard coded timeouts. FIXME(yang): deprecate this. Use the 21106 + errStatusMsg
• MODEL_TRAINING_WAITING_ERROR: Training got error waiting on asset pipeline to finish. FIXME(yang): deprecate this. Use the 21106 + errStatusMsg
• MODEL_TRAINING_UNKNOWN_ERROR: Training threw an unknown exception.
• MODEL_TRAINING_MSG_REDELIVER: Training message was redelivered. FIXME(yang): deprecate this. Use the 21106 + errStatusMsg
• MODEL_TRAINING_INSUFFICIENT_DATA: Training got error due to insufficient labelled data. FIXME(yang): deprecate this. Use the 21106 + errStatusMsg
• MODEL_TRAINING_INVALID_PARAMS: FIXME(yang): deprecate this. Use the 21106 + errStatusMsg
• MODEL_TRAINING_INVALID_DATA_TOLERANCE_EXCEEDED: Training is stopped because too much data was dropped. FIXME(yang): deprecate this. Use the 21106 + errStatusMsg
• MODEL_EVALUATION_TIMED_OUT: Evaluation took longer than hard coded timeouts. FIXME(yang): deprecate this. Use the 21317 + errStatusMsg
• MODEL_EVALUATION_WAITING_ERROR: Evaluation got error waiting on asset pipeline to finish.FIXME(yang): deprecate this. Use the 21317 + errStatusMsg
• MODEL_EVALUATION_UNKNOWN_ERROR: EVALUATION THREW AN UNKNOWN EXCEPTION.
• MODEL_EVALUATION_MSG_REDELIVER: Eval message was redelivered. FIXME(yang): deprecate this. Use the 21317 + errStatusMsg
• MODEL_EVALUATION_NEED_LABELS: Don’t have enough concepts labelled to perform evaluation. FIXME(yang): deprecate this. Use the 21317 + errStatusMsg
• MODEL_EVALUATION_NEED_INPUTS: Don’t have enough inputs per concept to perform evaluation. FIXME(yang): deprecate this. Use the 21317 + errStatusMsg
• MODEL_EVALUATION_FAILED: Generic err code for eval failure.
• MODEL_DEPLOYMENT_FAILED: Used when inference coordinator failed to deploy spire and throws an error
• MODEL_DEPLOYING: Used when calling the inference coordinator to deploy a spire
• MODEL_QUEUED_FOR_DEPLOYMENT: Used when training is completed
• MODEL_NOT_DEPLOYED: Used when model spire deployment is manually taken down or due to inactivity
• MODEL_REFERENCE_INVALID_ARGUMENT: Used when a model reference field is not set properly
• MODEL_EXAMPLE_INPUT_INVALID_ARGUMENT: Used when a model example input field is not set properly
• MODEL_EXPORTED: Model Export status codes
• WORKFLOW_NO_MATCHING_INPUT: specified model input not in workflow
• WORKFLOW_REQUIRE_TRAINED_MODEL: specified model must be trained
• WORKFLOW_INVALID_ARGUMENT: error in the request somewhere
• WORKFLOW_INVALID_REQUEST: error in the request somewhere
• CONCEPT_MODIFY_SUCCESS: Concept related 23xxx
• ANNOTATION_SUCCESS: Annotation related 24xxx
• METADATA_INVALID_PATCH_ARGUMENTS: Metadata related 249xx
• TRAINER_JOB_STATE_NONE: Training service related 25xxx
• DATA_DUMP_SUCCESS: Data Dump related 251xx
• DATA_DUMP_NO_DATA: DEPRECATED: Not used anymore. Now for an empty data dump, DATA_DUMP_SUCCESS is returned. To detect an empty data dump, check if the inptus count is 0.
• APP_DUPLICATION_SUCCESS: Duplicate related 252xx
• MODULE_DOES_NOT_EXIST: Module related codes 253xx
• BULK_OPERATION_SUCCESS: Bulk Operation related codes 254xx
• RUNNER_DOES_NOT_EXIST: Runner related codes 256xx
• INPUT_DOWNLOAD_SUCCESS: Input:Image related 30xxx
• INPUT_DOWNLOAD_PENDING: when things are async, this is the default status.
• INPUT_DOWNLOAD_FAILED: any type of error downloading and processing
• INPUT_VIDEO_DOWNLOAD_SUCCESS: Input:Video related 31xxx – Deprecated
• INPUT_WRITES_DISABLED_FOR_MAINTENANCE: deprecate this one. Use REQUEST_DISABLED_FOR_MAINTENANCE
• PREDICT_INVALID_REQUEST: API formatting issues 4000x
• DATABASE_DUPLICATE_KEY: Other related 400xx
• EXTERNAL_CONNECTION_ERROR: could not connect to external services
• QUEUE_CONN_ERROR: Queue related errors 41xxx
• SQS_OVERLIMIT: SQS related errors 411xx
• SEARCH_INTERNAL_FAILURE: Search related errors 43xxxx
• EVALUATION_QUEUED: Workflow evaluation err code
• STRIPE_EVENT_ERROR: Stripe 44xxx
• CACHE_MISS: Redis/Cache 45xxx
• SIGNUP_EVENT_ERROR: Sift Science 46xxx
• APP_COUNT_INVALID_MESSAGE: Application counts related errors 470xx
• MP_DOWNLOAD_ERROR: Media processor related errors 471xx – DEPRECATED
• DATATIER_CONN_ERROR: DataTier related error 472xx
• USER_CONSENT_FACE: User legal consent stauts related 50xxx
• WORKER_MISSING: Workers 51xxx
• COLLECTOR_MISSING: Collectors 52xxx
• SSO_IDENTITY_PROVIDER_DOES_NOT_EXIST: SSO 53xxx
• TASK_IN_PROGRESS: Tasks 54xxx
  The task was created.
• TASK_DONE: The task is completed.
• TASK_WONT_DO: The task is marked as abandoned.
• TASK_FAILED: An error occurred during add-task-annotations or add-auto-annotations pipeline.
• TASK_IDLE: When an Auto Annotation task job has finished processing its last batch and is waiting for more dataset assets.
• TASK_CONFLICT: The task operation is in conflict with the current state of the server.
• TASK_NOT_IMPLEMENTED: Certain task-related scenarios are not implemented.
• TASK_MISSING: Task was not found.
• TASK_PERMISSION_DENIED: Not allowed to perform a task-related action.
• LABEL_ORDER_PENDING: Label Order Related Status Code 55xxx
• LICENSE_ACTIVE: License Related Status Code 600xx
• LICENSE_DELETED: hidden state not reflected to users
• PASSWORD_VALIDATION_SUCCESS: Password Related Status Code
• FEATUREFLAG_CONFIG_NOT_FOUND: Feature flags status code
• MAINTENANCE_SUCCESS: Maintenance status code
• DATASET_VERSION_PENDING: Datasets 64xxx
  The dataset version is pending to be processed.
• DATASET_VERSION_IN_PROGRESS: The dataset version is currently being processed.
• DATASET_VERSION_READY: The dataset version is ready to be used.
• DATASET_VERSION_FAILURE: An error occurred during the dataset version processing.
• DATASET_VERSION_UNEXPECTED_ERROR: An unexpected error occurred during the dataset version processing.
• DATASET_VERSION_CONFLICT: An alteration to dataset version would create a conflict
• DATASET_INPUT_SUCCESS: The dataset input was successfully added.
• DATASET_INPUT_DUPLICATE: The dataset input is a duplicate.
  Deprecated: Unused.
• DATASET_VERSION_EXPORT_SUCCESS: The dataset version export is completed.
• DATASET_VERSION_EXPORT_PENDING: The dataset version is pending to be exported.
• DATASET_VERSION_EXPORT_FAILED: An error occurred during the dataset version export.
• DATASET_VERSION_EXPORT_IN_PROGRESS: The dataset version is currently being exported.
• DATASET_VERSION_EXPORT_UNEXPECTED_ERROR: An unexpected error occurred during the dataset version export.
• JOB_QUEUED: Generic Job status codes
• AUTH_MISSING_IDP_ASSOC: auth issues

TODO: Knowledge graph related 80xxx

• UPLOAD_IN_PROGRESS: Multipart uploading status codes
• BILLING_INVALID_INFO: Billing related issues: 69xxx
• INTERNAL_SERVER_ISSUE: Internal issues: 98xxx
• CONN_UNCATEGORIZED: Uncategorized: 99xxx: move off as soon as known
• BAD_REQUEST: Deprecated: migrate off to one of the internal issues
• SERVER_ERROR: Deprecated: migrate off to one of the internal issues

A short description of the error.

More details of the given error.
These details may be exposed to non-technical users.
For technical details, try to use developer_notes field.

For some environment we may return a stack trace to help debug
any issues.

specifically for long running jobs

if status is pending, how much time is remaining (in seconds)

A request ID may be present, to help monitoring and tracking requests

Internal Annotation (do not set in production, for internal Clarifai use only).

New location for the resource. Used to set response Location header.

Resource type

Old resource id

New resource id

Notes for developer.
These notes are rather technical details for developers how to interpret the status,
e.g. why an error occurred and how to avoid getting the error.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

id of add inputs job

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 20.

By default, the endpoint return inputs by the time when it is added.
If this is set to true, we will return inputs by id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

(optional URL parameter) List of additional fields to be included in the response. Currently supported: all, stars, outputs, presets

Sorting options:
Whether to sort in ascending order. If false, will order in descending order.

Whether to order by the name

Whether to order by the number of training inputs

Whether to order by the modified_at time of the latest model version.
If none of the sort options is set to true, will sort by modified_at.

Whether to order by the created_at

Whether to order by count of stars

Filtering options:
Filter models by the specific model_type_id. See ListModelTypes for the list of ModelType.Id’s
supported.

If true, we only return models that have the status MODEL_TRAINED, which includes non-trainable model types.

The list of input fields to the model.
For example, you can specify ‘image’, which will return models that make inferences on images like visual-classifier models.

The list of output fields to the model.
For example, you can specify ‘regions[…].data.concepts’, which will return visual-detector models.

Filter by the license of the model version

If true, we only return models that are handpicked by clarifai staff

If true, we only return models that are starred by the requesting user

List of toolkit tags to filter by

List of use_case tags to filter by

List of language tags to filter by

Old API behavior resulted in returning clarifai main models when calling ListModels while scoped to an app. While we transition
away from that, we can use this flag to not always fetch clarifai main models, unless that is the app we are explicitly listing for.

Filter models by bookmark. If set, only return bookmarked models. Otherwise none bookmarked models only.
Note: you can not filter trained_only and bookmark at the same time.
When filter by bookmark, we will return trained and untrained models.

Searching options:
Specify a search parameter in order to perform keyword search on the
following fields of the model:

• id
• name
• description
• notes
• user_id (unless user_app_id.user_id is already set)

Keywords are both normalized for search (so searching for “satisfy” matches “satisfied”)
and used for partial prefix-matching (so searching for “clari” matches “clarifai”).

NOTE: Both the list of fields searched and the exact keyword matching
rules are subject to change and not guaranteed to be backwards-compatible.

Query name, description and id fields, that can contain the words in the query string. Does NOT support wildcards - full words only. Supports operators “OR” and “-“ as NOT.
Deprecated: use search instead.

Filter by the description and id of the model. This supports wildcard queries like “gen*” to match “general” as an example.
Deprecated: use search instead.

Extends the name filter to include the user_id of the application owner that the model belongs to.
Deprecated: use search instead of name.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

The specific ModelType.Id you want to retrieve.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

This is included so that we can re-use this request for multiple
rpcs with and without the version_id.

(optional URL parameter) List of additional fields to be included in the response. Currently supported: all, stars, outputs, presets

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

This is included so that we can re-use this request for multiple
rpcs with and without the version_id.

(optional URL parameter) List of additional fields to be included in the response. Currently supported: all, stars, outputs, presets

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Optional, defaults to 1.

Optional, defaults to 128 references per page.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Filtering options:
To list only the model versions that have these concept ids present in them.

To list only the model versions that have been trained.

Sorting options:
Whether to sort in ascending order. If false, will order in descending order.

Whether to order by the status code

Whether to order by the number of training inputs

Whether to sort by the description

Whether to order by the created_at time
If neither sort option is set to true, will sort by created_at.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

id of model the example belongs to

specific version the example belongs to

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Optional, defaults to 1.

Optional, defaults to 128 references per page.

id of model the example belongs to

specific version the example belongs to

Id of example to fetch

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

This is included so that we can re-use this request for multiple
rpcs with and without the version_id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) List of additional fields to be included in the response. Currently supported: all, stars, outputs, presets

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Sorting options:
Whether to sort in ascending order. If false, will order in descending order.

Whether to order by the created_at time.

Whether to order by the number of users stared the app

If neither sort option is set to true, will sort by modified_at.

Whether to order by the external id

Filtering options:

Filter modules by bookmark. If set, only return bookmarked modules. Otherwise none bookmarked modules only.

Searching options:
Specify a search parameter in order to perform keyword search on the
following fields of the module:

• id
• description
• user_id (unless user_app_id.user_id is already set)

Keywords are both normalized for search (so searching for “satisfy” matches “satisfied”)
and used for partial prefix-matching (so searching for “clari” matches “clarifai”).

NOTE: Both the list of fields searched and the exact keyword matching
rules are subject to change and not guaranteed to be backwards-compatible.

Filter by the id and description of the module. This supports wildcard queries like “gen*” to match “general” as an example.
Deprecated: use search instead.

Filter by the application owner whose this module belongs to
Deprecated: use search instead of name.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Get tasks that have ANY user from this list assigned as worker.

Get tasks that have ANY user from this list assigned as reviewer.

Get tasks that are associated to ANY label order from this list.

Get label order tasks as well
It is automatically set to true if label_order_ids is set.

(optional URL parameter) List of additional fields to be included in the response.
Currently supported additional fields:

• all
• worker.users
• review.users
• metrics.work.inputs_count_estimated
• metrics.work.inputs_percent_estimated

(optional) task IDs to filter on

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) List of additional fields to be included in the response.
Currently supported additional fields:

• all
• worker.users
• review.users
• metrics.work.inputs_count_estimated
• metrics.work.inputs_percent_estimated

task_id for which count per user per status is needed

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

for given task_id, user_ids to filter on (optional)

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

task_id for which count per user per status is needed

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

for given task_id, user_ids to filter on (optional)

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

URL param. If zero ids provided, returns for all task labelers

For now view types ‘apps’, ‘workflows’, and ‘models’ are supported.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Upload ID

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) List of additional fields to be included in the response. Currently supported: all, stars

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

(optional URL parameter) List of additional fields to be included in the response. Currently supported: all, stars

Sorting options:
Whether to sort in ascending order. If false, will order in descending order.

Whether to order by the name

Whether to order by the modified_at time.
If none of the sort options is set to true, will sort by modified_at.

Whether to order by the created_at time.

Whether to order by the number of users stared the app

Filtering options:
If true, we only return apps that are handpicked by clarifai staff

If true, we only return apps that are starred by the requesting user

If true, we only return apps that are marked as a template by the app owner.

Searching options:
Specify a search parameter in order to perform keyword search on the
following fields of the application:

• id
• name
• description
• notes
• user_id (unless user_app_id.user_id is already set)

Keywords are both normalized for search (so searching for “satisfy” matches “satisfied”)
and used for partial prefix-matching (so searching for “clari” matches “clarifai”).

NOTE: Both the list of fields searched and the exact keyword matching
rules are subject to change and not guaranteed to be backwards-compatible.

Query various text fields (id, name, description, and notes) that can contain the words in the query string
Deprecated: use search instead.

Filter by the id, name and notes of the app. This supports wilcard queries like “gen*” to match “general” as an example.
Deprecated: use search instead.

Filter by the user-unique-id of the app. This supports wilcard queries like “gen*” to match “general” as an example.
Deprecated: use search instead.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) List of additional fields to be included in the response. Currently supported: all, stars

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Identify annotation filter by id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

List annotations for these IDs

List annotations for these Input IDs
Note that ‘ids’ are optional but if the are provided the number and order in
‘ids’ and ‘input_ids’ should match
If you do not specify ‘ids’ all the annotations for ‘input_ids’ are returned
If a you do not specify both ‘input_ids’ and ‘ids’ all the annotations in the app are returned

Only return the annotations that has one of these user IDs, effectively operating as an
OR among them to filter down the results.
If model_version_ids are also provided these user_ids are OR’d with them as well since
annotations are either provided by users or model versions and we want the union of any
provided user or model version annotations in the results of ListAnnotations request.
If no user_ids are provided then annotations from all users are returned.

Only return the annotations that has one of these model version IDs, effectively operating as an
OR among them to filter down the results.
If user_ids are also provided these model_versions_ids are OR’d with them as well since
annotations are either provided by users or model versions and we want the union of any
provided user or model version annotations in the results of ListAnnotations request.
If no model_version_ids are provided then annotations from all model versions are returned.

Set this flag to list both trusted and not trusted annotations
by default it’s listing only trusted annotations

Set this flag to return the model output for model annotations in the response.

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Flag to filter annotations by task_id

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Unique custom facing id that identifies the eval to get

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

(optional URL parameter) List of additional fields to be included in the response.
Currently supported: all, names

(optional URL parameter) Only list workers that have created trusted annotations.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

The bulk operation id

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Set this flag to list both deleted and not deleted collaborators
by default it’s listing only not deleted collaborators

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Fuzzy match on concept ID

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

The subject concept id in your app to get all the relationships for.
Leave as an empty string (GET /concepts/relations) to list ALL the relations in the app.

When listing all the relations it will only return one direction of the relationship
with the predicate acting on the subject and not the inverse like is done when providing a
concept_id so that we can return a reliable page size always.

When providing a concept_id, if a hyponym is present in the DB such as:
‘honey’ (subject), ‘hyponym’ (predicate for “is a kind of”), ‘food’ (object)
then you can list the concept relations for ‘honey’ and get hyponym predicate with ‘food’
object.
But you can also list the concept relations for ‘food’ and it will return the same hyponym
relationship with ‘honey’ as object and ‘hypernym’ as predicate.
Synonyms by nature are symmetrical relationships so either side can be the concept_id (subject)
when listing the relations.

If predicate is provided then only list relations with that predicate.

Note that if no subject is set in concept_id and predicate is set to
‘hypernym’, then it will return any stored hyponyms as hypernyms with
just the subject and object swapped since they are reversed relations.

Valid predicates are:

• ‘hypernym’
• ‘hyponym’
• ‘synonym’

If knowledge_graph_id is provided then just list relations from that knowledge graph.
If not provided then list relations from all knowledge graphs including the global one for this
app one (ie. knowledge_graph “”) and any specific ones in the app.

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

The concept’s id.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

(optional URL parameter) The page number. Pagination is used to split the results into chunks.
Defaults to 1.

(optional URL parameter) The number of results that will be contained in each page. Defaults
to 128.

Note user_id ‘me’ is reserved - it is the alias for the id of authorized user

This is the language identifier.