Label Studio
App · 19 actions
Label data for machine learning in Label Studio: create projects, import tasks, add annotations and model predictions, and export labels. Start a workflow when a task is annotated or labelled, or a project is finished.
Actions
| Action | What it does |
|---|---|
| Add prediction | Add a model's prediction to a task, for annotators to start from: a plain answer such as Positive, or Label Studio's result format, with a score. |
| Add predictions to tasks | Add a model's predictions to many tasks of a project at once, each a task ID with an answer or a result, and a score. |
| Create annotation | Label a task: add an annotation from a plain answer, such as a person's choice of Positive, or from Label Studio's result format. |
| Create project | Create a project from a ready setup, such as text classification with your labels, or from a labelling config of your own. |
| Delete annotation | Delete an annotation for good. |
| Delete project | Delete a project for good, with its tasks, annotations and predictions. |
| Delete task | Delete a task for good, with its annotations and predictions. |
| Export annotations | Export a project's tasks with their annotations as JSON, the labelled ones or all, such as to train a model or back them up. |
| Get annotation | Get an annotation with the labels it picks, who made it and when. |
| Get project | Get a project with its progress: how many tasks it has, how many are labelled, and whether it is finished. |
| Get task | Get a task with its data, whether it is labelled, and its annotations and predictions. |
| Import tasks | Add tasks to a project from rows or JSON: each object is one task's data, such as a text to label, or a whole Label Studio task. |
| Import tasks from links | Add a task for each link to a file to label, such as an image or audio file. |
| List annotations | List a task's annotations, with the labels each one picks. |
| List members | List the organization's members, with the IDs annotations name them by. |
| List projects | List projects, newest first, with how many of each one's tasks are labelled. |
| List tasks | List a project's tasks, all of them or only the labelled or unlabelled ones, with their data. |
| Update annotation | Change an annotation's labels, from a plain answer or Label Studio's result format, or mark it as ground truth. |
| Update project | Change a project's title, description, labelling config, instructions or annotations per task. |
Add prediction
Add a model's prediction to a task, for annotators to start from: a plain answer such as Positive, or Label Studio's result format, with a score.
Inputs
- Task ID
task_idtextRequired - The task's ID, a number shown in the project's task list, or a link to the task.
- Answer
answertext - A plain answer for the project's choices, text box, rating or number: a choice such as Positive (several separated by commas), the text, or a number. Leave blank to give a result instead.
- Result
resultlist - The labels in Label Studio's own result format, for anything a plain answer can't say (a span of text, a box on an image): a list of regions, each with from_name, to_name, type and value. Leave blank when giving an answer.
- Control
controltext - Which control of the labelling config the answer is for, by its name, such as sentiment. Only needed when the config has more than one.
- Score
scorenumber - The model's confidence, from 0 to 1, such as 0.87.From 0 to 1.
- Model version
model_versiontext - The name of the model version the predictions came from, such as sentiment-2026-11.
Outputs
created_attext- When it was added (ISO 8601)
labelslist- The choices and labels it picks
model_versiontext- The model version it came from
prediction_idnumber- The prediction's ID
resultlist- The labels as Label Studio stores them: regions with from_name, to_name, type and value
scorenumber- The model's confidence, from 0 to 1
task_idnumber- The task it is for
Example output
{
"created_at": "2026-11-10T13:58:12.004918Z",
"labels": [
"Positive"
],
"model_version": "sentiment-2026-11",
"prediction_id": 301,
"result": [
{
"from_name": "label",
"to_name": "text",
"type": "choices",
"value": {
"choices": [
"Positive"
]
}
}
],
"score": 0.91,
"task_id": 412
}Add predictions to tasks
Add a model's predictions to many tasks of a project at once, each a task ID with an answer or a result, and a score.
Inputs
- Project
projecttextRequired - The project. Pick one, or give its ID or a link to it.
- Predictions
predictionslistRequired - A list of predictions, such as [{"task": 412, "answer": "Positive", "score": 0.91}]: each names its task by ID and gives an answer, or a result in Label Studio's format, and may give a score from 0 to 1.
- Model version
model_versiontext - The name of the model version the predictions came from, such as sentiment-2026-11.
- Control
controltext - Which control of the labelling config the answers are for. Only needed when the config has more than one.
Outputs
creatednumber- How many predictions were added
model_versiontext- The model version given for all of them
project_idnumber- The project
task_idslist- The tasks they were added to
Example output
{
"created": 2,
"model_version": "sentiment-2026-11",
"project_id": 3,
"task_ids": [
412,
413
]
}Create annotation
Label a task: add an annotation from a plain answer, such as a person's choice of Positive, or from Label Studio's result format.
Inputs
- Task ID
task_idtextRequired - The task's ID, a number shown in the project's task list, or a link to the task.
- Answer
answertext - A plain answer for the project's choices, text box, rating or number: a choice such as Positive (several separated by commas), the text, or a number. Leave blank to give a result instead.
- Result
resultlist - The labels in Label Studio's own result format, for anything a plain answer can't say (a span of text, a box on an image): a list of regions, each with from_name, to_name, type and value. Leave blank when giving an answer.
- Control
controltext - Which control of the labelling config the answer is for, by its name, such as sentiment. Only needed when the config has more than one.
- Ground truth
ground_truthtrue/false - Mark the annotation as ground truth, the answer others are compared with.
- Time taken
lead_time_secondsnumber - How long labelling took, in seconds, for Label Studio's statistics.At least 0.
Outputs
annotation_idnumber- The annotation's ID
annotator_idnumber- The ID of the member who made it
answersobject- What each control was answered, by its name: a list of choices or labels, text, a rating or a number
created_attext- When it was made (ISO 8601)
ground_truthtrue/false- True when it is marked as ground truth
labelslist- The choices and labels it picks, such as Positive
lead_time_secondsnumber- How long the annotator took, in seconds
linktext- It in Label Studio
project_idnumber- The task's project
resultlist- The labels as Label Studio stores them: regions with from_name, to_name, type and value
skippedtrue/false- True when the annotator skipped the task
task_idnumber- The task it labels
updated_attext- When it last changed (ISO 8601)
Example output
{
"annotation_id": 88,
"annotator_id": 2,
"answers": {
"label": [
"Positive"
]
},
"created_at": "2026-11-10T14:02:51.118404Z",
"ground_truth": false,
"labels": [
"Positive"
],
"lead_time_seconds": 6.4,
"link": "https://label-studio.example.com/projects/3/data?task=412&annotation=88",
"project_id": 3,
"result": [
{
"from_name": "label",
"to_name": "text",
"type": "choices",
"value": {
"choices": [
"Positive"
]
}
}
],
"skipped": false,
"task_id": 412,
"updated_at": "2026-11-10T14:02:51.118420Z"
}Create project
Create a project from a ready setup, such as text classification with your labels, or from a labelling config of your own.
Inputs
- Title
titletextRequired - The project's name, up to 50 characters.
- Setup
templatetext - What annotators do. Defaults to text classification.One of:
text_classification(Text classification: pick a choice for a text),named_entities(Named entities: mark spans of a text),image_classification(Image classification: pick a choice for an image),object_detection(Object detection: draw boxes on an image),custom(My own labelling config). Default:text_classification. - Labels
labelslist - The choices or labels to offer, such as Positive and Negative. Needed for a ready setup.
- Labelling config
label_configtext - A labelling config of your own in Label Studio's XML, starting with <View>, as in a project's Settings > Labeling Interface > Code. Used instead of the setup.
- Description
descriptiontext - What the project is for.
- Instructions
instructionstext - Instructions shown to annotators before they start.
- Annotations per task
annotations_per_tasknumber - How many people annotate each task before it counts as labelled. Defaults to 1.From 1 to 100.
Outputs
annotation_countnumber- How many annotations it has, skipped ones included
annotations_per_tasknumber- How many annotations make a task labelled
created_attext- When it was made (ISO 8601)
created_bytext- The email of whoever made it
data_fieldslist- The task data the config reads, such as text or image
descriptiontext- Its description
finishedtrue/false- True when it has tasks and every one is labelled
instructionstext- The instructions shown to annotators
label_configtext- Its labelling config, in Label Studio's XML
labelled_task_countnumber- How many of them are labelled
labelslist- Every choice and label the config offers
linktext- Its tasks in Label Studio
percent_labellednumber- The share labelled, from 0 to 100
prediction_countnumber- How many predictions it has
project_idnumber- The project's ID
skipped_countnumber- How many annotations were skipped
task_countnumber- How many tasks it has
titletext- Its name
unlabelled_task_countnumber- How many are still to label
Example output
{
"annotation_count": 47,
"annotations_per_task": 1,
"created_at": "2026-11-02T09:14:05.481236Z",
"created_by": "ana@example.com",
"data_fields": [
"text"
],
"description": "Is each review positive or negative?",
"finished": false,
"instructions": "",
"label_config": "<View>\n <Text name=\"text\" value=\"$text\"/>\n <Choices name=\"label\" toName=\"text\" choice=\"single\" showInline=\"true\">\n <Choice value=\"Positive\"/>\n <Choice value=\"Negative\"/>\n </Choices>\n</View>",
"labelled_task_count": 45,
"labels": [
"Positive",
"Negative"
],
"link": "https://label-studio.example.com/projects/3/data",
"percent_labelled": 37.5,
"prediction_count": 120,
"project_id": 3,
"skipped_count": 2,
"task_count": 120,
"title": "Review sentiment",
"unlabelled_task_count": 75
}Delete annotation
Delete an annotation for good. The task stays.
Inputs
- Annotation ID
annotation_idtextRequired - The annotation's ID, a number, or a link to it.
Outputs
annotation_idnumber- The deleted annotation's ID
deletedtrue/false- True once the annotation is deleted
task_idnumber- The task it was on
Example output
{
"annotation_id": 88,
"deleted": true,
"task_id": 412
}Delete project
Delete a project for good, with its tasks, annotations and predictions.
Inputs
- Project
projecttextRequired - The project. Pick one, or give its ID or a link to it.
Outputs
deletedtrue/false- True once the project is deleted
project_idnumber- The deleted project's ID
task_countnumber- How many tasks were deleted with it
titletext- Its name
Example output
{
"deleted": true,
"project_id": 3,
"task_count": 120,
"title": "Review sentiment"
}Delete task
Delete a task for good, with its annotations and predictions.
Inputs
- Task ID
task_idtextRequired - The task's ID, a number shown in the project's task list, or a link to the task.
Outputs
deletedtrue/false- True once the task is deleted
task_idnumber- The deleted task's ID
Example output
{
"deleted": true,
"task_id": 412
}Export annotations
Export a project's tasks with their annotations as JSON, the labelled ones or all, such as to train a model or back them up.
Inputs
- Project
projecttextRequired - The project. Pick one, or give its ID or a link to it.
- Format
export_formattext - JSON gives each task with its data, annotations and predictions; JSON-MIN gives one simple row per annotation. Defaults to JSON.One of:
JSON(JSON (tasks with annotations)),JSON_MIN(JSON-MIN (one row per annotation)). Default:JSON. - Only labelled tasks
only_labelledtrue/false - Export only tasks with annotations. Defaults to on; turn it off for every task.Default:
true.
Outputs
export_formattext- JSON or JSON_MIN
only_labelledtrue/false- True when only labelled tasks were exported
project_idnumber- The project
task_countnumber- How many tasks were exported
taskslist- The tasks: with JSON, each with its data, annotations and predictions as Label Studio exports them; with JSON_MIN, one row per annotation with the data and each control's answer
Example output
{
"export_format": "JSON_MIN",
"only_labelled": true,
"project_id": 3,
"task_count": 1,
"tasks": [
{
"annotation_id": 88,
"annotator": 2,
"created_at": "2026-11-10T14:02:51.118404Z",
"id": 412,
"label": "Positive",
"lead_time": 6.4,
"text": "Arrived quickly and works perfectly.",
"updated_at": "2026-11-10T14:02:51.118420Z"
}
]
}Get annotation
Get an annotation with the labels it picks, who made it and when.
Inputs
- Annotation ID
annotation_idtextRequired - The annotation's ID, a number, or a link to it.
Outputs
annotation_idnumber- The annotation's ID
annotator_idnumber- The ID of the member who made it
answersobject- What each control was answered, by its name: a list of choices or labels, text, a rating or a number
created_attext- When it was made (ISO 8601)
ground_truthtrue/false- True when it is marked as ground truth
labelslist- The choices and labels it picks, such as Positive
lead_time_secondsnumber- How long the annotator took, in seconds
linktext- It in Label Studio
project_idnumber- The task's project
resultlist- The labels as Label Studio stores them: regions with from_name, to_name, type and value
skippedtrue/false- True when the annotator skipped the task
task_idnumber- The task it labels
updated_attext- When it last changed (ISO 8601)
Example output
{
"annotation_id": 88,
"annotator_id": 2,
"answers": {
"label": [
"Positive"
]
},
"created_at": "2026-11-10T14:02:51.118404Z",
"ground_truth": false,
"labels": [
"Positive"
],
"lead_time_seconds": 6.4,
"link": "https://label-studio.example.com/projects/3/data?task=412&annotation=88",
"project_id": 3,
"result": [
{
"from_name": "label",
"to_name": "text",
"type": "choices",
"value": {
"choices": [
"Positive"
]
}
}
],
"skipped": false,
"task_id": 412,
"updated_at": "2026-11-10T14:02:51.118420Z"
}Get project
Get a project with its progress: how many tasks it has, how many are labelled, and whether it is finished.
Inputs
- Project
projecttextRequired - The project. Pick one, or give its ID or a link to it.
Outputs
annotation_countnumber- How many annotations it has, skipped ones included
annotations_per_tasknumber- How many annotations make a task labelled
created_attext- When it was made (ISO 8601)
created_bytext- The email of whoever made it
data_fieldslist- The task data the config reads, such as text or image
descriptiontext- Its description
finishedtrue/false- True when it has tasks and every one is labelled
instructionstext- The instructions shown to annotators
label_configtext- Its labelling config, in Label Studio's XML
labelled_task_countnumber- How many of them are labelled
labelslist- Every choice and label the config offers
linktext- Its tasks in Label Studio
percent_labellednumber- The share labelled, from 0 to 100
prediction_countnumber- How many predictions it has
project_idnumber- The project's ID
skipped_countnumber- How many annotations were skipped
task_countnumber- How many tasks it has
titletext- Its name
unlabelled_task_countnumber- How many are still to label
Example output
{
"annotation_count": 47,
"annotations_per_task": 1,
"created_at": "2026-11-02T09:14:05.481236Z",
"created_by": "ana@example.com",
"data_fields": [
"text"
],
"description": "Is each review positive or negative?",
"finished": false,
"instructions": "",
"label_config": "<View>\n <Text name=\"text\" value=\"$text\"/>\n <Choices name=\"label\" toName=\"text\" choice=\"single\" showInline=\"true\">\n <Choice value=\"Positive\"/>\n <Choice value=\"Negative\"/>\n </Choices>\n</View>",
"labelled_task_count": 45,
"labels": [
"Positive",
"Negative"
],
"link": "https://label-studio.example.com/projects/3/data",
"percent_labelled": 37.5,
"prediction_count": 120,
"project_id": 3,
"skipped_count": 2,
"task_count": 120,
"title": "Review sentiment",
"unlabelled_task_count": 75
}Get task
Get a task with its data, whether it is labelled, and its annotations and predictions.
Inputs
- Task ID
task_idtextRequired - The task's ID, a number shown in the project's task list, or a link to the task.
Outputs
annotation_countnumber- How many annotations it has
annotationslist- Its annotations
completed_attext- When it was last annotated (ISO 8601); blank if never
created_attext- When it was imported (ISO 8601)
dataobject- Its data, such as the text or image link to label
labelledtrue/false- True once it has the annotations the project asks for
linktext- It in Label Studio
prediction_countnumber- How many predictions it has
predictionslist- Its predictions
project_idnumber- Its project
skipped_countnumber- How many times it was skipped
task_idnumber- The task's ID
updated_attext- When it last changed (ISO 8601)
Example output
{
"annotation_count": 1,
"annotations": [
{
"annotation_id": 88,
"annotator_id": 2,
"answers": {
"label": [
"Positive"
]
},
"created_at": "2026-11-10T14:02:51.118404Z",
"ground_truth": false,
"labels": [
"Positive"
],
"lead_time_seconds": 6.4,
"link": "https://label-studio.example.com/projects/3/data?task=412&annotation=88",
"project_id": 3,
"result": [
{
"from_name": "label",
"to_name": "text",
"type": "choices",
"value": {
"choices": [
"Positive"
]
}
}
],
"skipped": false,
"task_id": 412,
"updated_at": "2026-11-10T14:02:51.118420Z"
}
],
"completed_at": "2026-11-10T14:02:51.118404Z",
"created_at": "2026-11-09T08:30:00.112233Z",
"data": {
"text": "Arrived quickly and works perfectly."
},
"labelled": true,
"link": "https://label-studio.example.com/projects/3/data?task=412",
"prediction_count": 1,
"predictions": [
{
"created_at": "2026-11-10T13:58:12.004918Z",
"labels": [
"Positive"
],
"model_version": "sentiment-2026-11",
"prediction_id": 301,
"result": [
{
"from_name": "label",
"to_name": "text",
"type": "choices",
"value": {
"choices": [
"Positive"
]
}
}
],
"score": 0.91,
"task_id": 412
}
],
"project_id": 3,
"skipped_count": 0,
"task_id": 412,
"updated_at": "2026-11-10T14:02:51.220011Z"
}Import tasks
Add tasks to a project from rows or JSON: each object is one task's data, such as a text to label, or a whole Label Studio task.
Inputs
- Project
projecttextRequired - The project. Pick one, or give its ID or a link to it.
- Tasks
taskslistRequired - The tasks, a list of objects such as [{"text": "Great service"}], with the data names the labelling config reads, or rows from an earlier step. An object with a data object is a whole Label Studio task, with its annotations or predictions.
Outputs
annotation_countnumber- How many annotations came with them
prediction_countnumber- How many predictions came with them
project_idnumber- The project
task_countnumber- How many tasks were imported
task_idslist- The new tasks' IDs
Example output
{
"annotation_count": 0,
"prediction_count": 0,
"project_id": 3,
"task_count": 2,
"task_ids": [
412,
413
]
}Import tasks from links
Add a task for each link to a file to label, such as an image or audio file.
Inputs
- Project
projecttextRequired - The project. Pick one, or give its ID or a link to it.
- Links
urlslistRequired - Links to the files, such as https://example.com/photos/1.jpg, one task each.
- Data field
data_fieldtext - The task data each link goes in, such as image. Only needed when the labelling config reads more than one.
Outputs
annotation_countnumber- How many annotations came with them
data_fieldtext- The task data each link went in, such as image
prediction_countnumber- How many predictions came with them
project_idnumber- The project
task_countnumber- How many tasks were imported
task_idslist- The new tasks' IDs
Example output
{
"annotation_count": 0,
"data_field": "image",
"prediction_count": 0,
"project_id": 3,
"task_count": 2,
"task_ids": [
412,
413
]
}List annotations
List a task's annotations, with the labels each one picks.
Inputs
- Task ID
task_idtextRequired - The task's ID, a number shown in the project's task list, or a link to the task.
Outputs
annotationslist- Its annotations, oldest first
countnumber- How many annotations it has
task_idnumber- The task
Example output
{
"annotations": [
{
"annotation_id": 88,
"annotator_id": 2,
"answers": {
"label": [
"Positive"
]
},
"created_at": "2026-11-10T14:02:51.118404Z",
"ground_truth": false,
"labels": [
"Positive"
],
"lead_time_seconds": 6.4,
"link": "https://label-studio.example.com/projects/3/data?task=412&annotation=88",
"project_id": 3,
"result": [
{
"from_name": "label",
"to_name": "text",
"type": "choices",
"value": {
"choices": [
"Positive"
]
}
}
],
"skipped": false,
"task_id": 412,
"updated_at": "2026-11-10T14:02:51.118420Z"
}
],
"count": 1,
"task_id": 412
}List members
List the organization's members, with the IDs annotations name them by.
Inputs
None.
Outputs
countnumber- How many members there are
memberslist- The organization's members
Example output
{
"count": 1,
"members": [
{
"email": "ana@example.com",
"last_activity": "2026-11-10T14:03:10.552211Z",
"name": "Ana Silva",
"user_id": 2,
"username": "ana"
}
]
}List projects
List projects, newest first, with how many of each one's tasks are labelled.
Inputs
- Title contains
searchtext - Only projects whose title contains this. Leave blank for every project.
- How many
limitnumber - The most projects to return, up to 1000. Defaults to 50.From 1 to 1000.
Outputs
countnumber- How many projects are listed
projectslist- The projects, newest first, with their progress
Example output
{
"count": 1,
"projects": [
{
"annotation_count": 47,
"annotations_per_task": 1,
"created_at": "2026-11-02T09:14:05.481236Z",
"created_by": "ana@example.com",
"data_fields": [
"text"
],
"description": "Is each review positive or negative?",
"finished": false,
"instructions": "",
"label_config": "<View>\n <Text name=\"text\" value=\"$text\"/>\n <Choices name=\"label\" toName=\"text\" choice=\"single\" showInline=\"true\">\n <Choice value=\"Positive\"/>\n <Choice value=\"Negative\"/>\n </Choices>\n</View>",
"labelled_task_count": 45,
"labels": [
"Positive",
"Negative"
],
"link": "https://label-studio.example.com/projects/3/data",
"percent_labelled": 37.5,
"prediction_count": 120,
"project_id": 3,
"skipped_count": 2,
"task_count": 120,
"title": "Review sentiment",
"unlabelled_task_count": 75
}
]
}List tasks
List a project's tasks, all of them or only the labelled or unlabelled ones, with their data.
Inputs
- Project
projecttextRequired - The project. Pick one, or give its ID or a link to it.
- Which tasks
statustext - All tasks, or only labelled or unlabelled ones. Defaults to all.One of:
all,labelled,unlabelled. Default:all. - Include annotations
include_annotationstrue/false - Also return each task's annotations and predictions.
- How many
limitnumber - The most tasks to return, up to 1000. Defaults to 50.From 1 to 1000.
Outputs
countnumber- How many are listed
project_idnumber- The project
statustext- Which tasks: all, labelled or unlabelled
taskslist- The tasks, oldest first
totalnumber- How many tasks of the project match
Example output
{
"count": 1,
"project_id": 3,
"status": "all",
"tasks": [
{
"annotation_count": 1,
"annotations": [
{
"annotation_id": 88,
"annotator_id": 2,
"answers": {
"label": [
"Positive"
]
},
"created_at": "2026-11-10T14:02:51.118404Z",
"ground_truth": false,
"labels": [
"Positive"
],
"lead_time_seconds": 6.4,
"link": "https://label-studio.example.com/projects/3/data?task=412&annotation=88",
"project_id": 3,
"result": [
{
"from_name": "label",
"to_name": "text",
"type": "choices",
"value": {
"choices": [
"Positive"
]
}
}
],
"skipped": false,
"task_id": 412,
"updated_at": "2026-11-10T14:02:51.118420Z"
}
],
"completed_at": "2026-11-10T14:02:51.118404Z",
"created_at": "2026-11-09T08:30:00.112233Z",
"data": {
"text": "Arrived quickly and works perfectly."
},
"labelled": true,
"link": "https://label-studio.example.com/projects/3/data?task=412",
"prediction_count": 1,
"predictions": [
{
"created_at": "2026-11-10T13:58:12.004918Z",
"labels": [
"Positive"
],
"model_version": "sentiment-2026-11",
"prediction_id": 301,
"result": [
{
"from_name": "label",
"to_name": "text",
"type": "choices",
"value": {
"choices": [
"Positive"
]
}
}
],
"score": 0.91,
"task_id": 412
}
],
"project_id": 3,
"skipped_count": 0,
"task_id": 412,
"updated_at": "2026-11-10T14:02:51.220011Z"
}
],
"total": 1
}Update annotation
Change an annotation's labels, from a plain answer or Label Studio's result format, or mark it as ground truth.
Inputs
- Annotation ID
annotation_idtextRequired - The annotation's ID, a number, or a link to it.
- Answer
answertext - A plain answer for the project's choices, text box, rating or number: a choice such as Positive (several separated by commas), the text, or a number. Leave blank to give a result instead.
- Result
resultlist - The labels in Label Studio's own result format, for anything a plain answer can't say (a span of text, a box on an image): a list of regions, each with from_name, to_name, type and value. Leave blank when giving an answer.
- Control
controltext - Which control of the labelling config the answer is for, by its name, such as sentiment. Only needed when the config has more than one.
- Ground truth
ground_truthtrue/false - Mark the annotation as ground truth, the answer others are compared with.
Outputs
annotation_idnumber- The annotation's ID
annotator_idnumber- The ID of the member who made it
answersobject- What each control was answered, by its name: a list of choices or labels, text, a rating or a number
created_attext- When it was made (ISO 8601)
ground_truthtrue/false- True when it is marked as ground truth
labelslist- The choices and labels it picks, such as Positive
lead_time_secondsnumber- How long the annotator took, in seconds
linktext- It in Label Studio
project_idnumber- The task's project
resultlist- The labels as Label Studio stores them: regions with from_name, to_name, type and value
skippedtrue/false- True when the annotator skipped the task
task_idnumber- The task it labels
updated_attext- When it last changed (ISO 8601)
Example output
{
"annotation_id": 88,
"annotator_id": 2,
"answers": {
"label": [
"Positive"
]
},
"created_at": "2026-11-10T14:02:51.118404Z",
"ground_truth": false,
"labels": [
"Positive"
],
"lead_time_seconds": 6.4,
"link": "https://label-studio.example.com/projects/3/data?task=412&annotation=88",
"project_id": 3,
"result": [
{
"from_name": "label",
"to_name": "text",
"type": "choices",
"value": {
"choices": [
"Positive"
]
}
}
],
"skipped": false,
"task_id": 412,
"updated_at": "2026-11-10T14:02:51.118420Z"
}Update project
Change a project's title, description, labelling config, instructions or annotations per task.
Inputs
- Project
projecttextRequired - The project. Pick one, or give its ID or a link to it.
- Title
titletext - The new name, up to 50 characters. Leave blank to keep it.
- Description
descriptiontext - The new description. Leave blank to keep it.
- Labelling config
label_configtext - A new labelling config in Label Studio's XML. Label Studio refuses one that drops labels annotations already use. Leave blank to keep it.
- Instructions
instructionstext - New instructions for annotators. Leave blank to keep them.
- Annotations per task
annotations_per_tasknumber - How many people annotate each task before it counts as labelled. Defaults to 1.From 1 to 100.
Outputs
annotation_countnumber- How many annotations it has, skipped ones included
annotations_per_tasknumber- How many annotations make a task labelled
created_attext- When it was made (ISO 8601)
created_bytext- The email of whoever made it
data_fieldslist- The task data the config reads, such as text or image
descriptiontext- Its description
finishedtrue/false- True when it has tasks and every one is labelled
instructionstext- The instructions shown to annotators
label_configtext- Its labelling config, in Label Studio's XML
labelled_task_countnumber- How many of them are labelled
labelslist- Every choice and label the config offers
linktext- Its tasks in Label Studio
percent_labellednumber- The share labelled, from 0 to 100
prediction_countnumber- How many predictions it has
project_idnumber- The project's ID
skipped_countnumber- How many annotations were skipped
task_countnumber- How many tasks it has
titletext- Its name
unlabelled_task_countnumber- How many are still to label
Example output
{
"annotation_count": 47,
"annotations_per_task": 1,
"created_at": "2026-11-02T09:14:05.481236Z",
"created_by": "ana@example.com",
"data_fields": [
"text"
],
"description": "Is each review positive or negative?",
"finished": false,
"instructions": "",
"label_config": "<View>\n <Text name=\"text\" value=\"$text\"/>\n <Choices name=\"label\" toName=\"text\" choice=\"single\" showInline=\"true\">\n <Choice value=\"Positive\"/>\n <Choice value=\"Negative\"/>\n </Choices>\n</View>",
"labelled_task_count": 45,
"labels": [
"Positive",
"Negative"
],
"link": "https://label-studio.example.com/projects/3/data",
"percent_labelled": 37.5,
"prediction_count": 120,
"project_id": 3,
"skipped_count": 2,
"task_count": 120,
"title": "Review sentiment",
"unlabelled_task_count": 75
}