Labelstud Label Studio API
Official vendor OpenAPI document · agent-readyAI/MLMl InferenceapiKey397 EndpointsREST
For Agents
Create labeling projects, load tasks, collect annotations, attach predictions, and export labeled data in Label Studio. Covers 397 endpoints across projects, tasks, annotations, and predictions.
Use for: Create a new labeling project for an image classification dataset, Load a batch of tasks into an existing project, Collect the annotations recorded on a specific task, Attach model predictions to tasks so annotators can review them
Not supported: Does not train models or run inference itself. Use for managing labeling projects, annotations, and dataset exports only.
The Label Studio API manages data labeling projects for machine learning teams, exposing projects, tasks, annotations, predictions, and exports over a REST interface. It lets you create a project, load tasks, collect human annotations, attach model predictions, and export the labeled dataset for training. Requests carry a token in the Authorization header and return structured JSON that a pipeline can act on directly.
Install Jentic One Beta
Connect the Label Studio API to your agent
Jentic One is a self-hosted execution layer for AI agents. It lets your agent call the Label Studio API, or any other public or private API you need. You set the rules, the agent never sees your credentials, and every call is logged.
Two steps, two machines. Install the instance in a safe environment, then register your agent from wherever it runs.
1
Step 1: Jentic One Host machine
# On the machine that will host your Jentic One instance:
curl -fsSL "https://jentic.com/install.sh?src=apis&api=%2Fapis%2Flabelstud.io%2Flabelstud" | sh2
Step 2: Agent machine
# On the machine where your agent runs (keep this separate from the instance):
curl -fsSL "https://jentic.com/install.sh?src=apis&api=%2Fapis%2Flabelstud.io%2Flabelstud" | sh
jentic register # connects your agent to your Jentic One instanceJentic One is in public beta. The setup above keeps your agent separate from the instance, which is what you want before using real credentials: an agent running as the same OS user as Jentic One can read its stored keys directly. Just evaluating? A single local install is fine to start. See the secure deployment guide for the tiers.
Capabilities
What an agent can do with Label Studio API.
Create and list labeling projects and read their task counts
List and create tasks within a project for annotators to label
Collect annotations on a task and update or delete them
List and create model predictions to pre-label tasks or measure agreement
Import tasks and prediction files into a project
Export a project's labeled data and retrieve export snapshots
Use Cases
Patterns agents use Label Studio API for, with concrete tasks.
★ AI Labeling Orchestrator via Jentic
An AI agent can drive a Label Studio labeling loop without a developer wiring the API by hand. Through Jentic the agent searches for the operation it needs by intent, receives the matching Label Studio endpoint and its input schema, and calls it with the token injected at execution time. This lets a data team delegate task loading, prediction upload, and export to an agent that works directly against live projects.
Search Jentic for 'import tasks into a project', load the operation schema, call POST /api/projects/{id}/import, then poll task counts
Model-Assisted Pre-Labeling
Teams that already have a model can pre-label tasks so annotators only correct instead of labeling from scratch. The Label Studio API creates predictions against tasks, and annotators then confirm or fix them, cutting the manual effort. This raises throughput on large datasets where the model is already reasonably accurate.
Call POST /api/predictions/ for each task with the model's output, then let annotators review them in the project
Training Data Export Pipeline
Once a project is annotated, an ML pipeline needs the labeled data in a training-ready format. The Label Studio API triggers and retrieves exports so a scheduled job can pull the latest labels whenever a project reaches a threshold. Because the export comes straight from the platform, the training set always reflects the current annotations.
Call POST /api/projects/{id}/exports/ to create an export, then GET the snapshot when it is ready
Key Endpoints
397 endpoints — the label studio api manages data labeling projects for machine learning teams, exposing projects, tasks, annotations, predictions, and exports over a rest interface.
METHOD
PATH
DESCRIPTION
GET
/api/projects/
List labeling projects
POST
/api/projects/
Create a project
GET
/api/tasks/
List tasks
POST
/api/tasks/{id}/annotations/
Create an annotation on a task
GET
/api/predictions/
List predictions
POST
/api/predictions/
Create a prediction
POST
/api/projects/{id}/import
Import tasks into a project
GET
/api/projects/{id}/export
Export a project's annotations
GET
/api/projects/
List labeling projects
POST
/api/projects/
Create a project
GET
/api/tasks/
List tasks
POST
/api/tasks/{id}/annotations/
Create an annotation on a task
GET
/api/predictions/
List predictions
POST
/api/predictions/
Create a prediction
POST
/api/projects/{id}/import
Import tasks into a project
GET
/api/projects/{id}/export
Export a project's annotations
Why Jentic?
What agents get from Jentic-routed access to this vendor.
Setup
Wiring the Label Studio API by hand means setting the Authorization token on every request and threading project, task, and export ids through the right endpoints yourself. Through Jentic you install once, import Label Studio from the API Directory, store the token once, and your agent calls it.
Permission scoping
Label Studio puts the record id in the URL path (/api/projects/{id}/, /api/tasks/{id}/), so a rule can pin your agent to reading projects and adding predictions. You choose the operations it may call, so destructive operations like deleting annotations are not included unless you add them.
Credential isolation
Your Label Studio token is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context.
Intent-based discovery
Agents search Jentic by intent such as 'import tasks into a project' or 'export labeled data', and Jentic returns the matching Label Studio operation with its input schema so the agent calls the right endpoint without browsing the reference docs.
Related APIs
Alternatives and complements available in the Jentic catalogue.
FAQs
Specific to using Label Studio API through Jentic.
What authentication does the Label Studio API use?
The Label Studio API authenticates with a token sent in the Authorization header, as declared in its OpenAPI spec. Through Jentic the token is stored encrypted by your own Jentic One instance and injected at execution time, so it never enters the agent's prompt, logs, or context.
Can I upload model predictions with the Label Studio API?
Yes. Call POST /api/predictions/ to attach a model's output to a task, so annotators review and correct pre-labels instead of labeling from scratch. This is the standard path for model-assisted labeling on large datasets.
What are the rate limits for the Label Studio API?
The OpenAPI spec does not specify rate limits. Check the Label Studio API documentation at https://labelstud.io/guide/ for current limits before running high-volume import or export jobs.
How do I export labeled data through Jentic?
Search Jentic for 'export project annotations', which resolves to the Label Studio export operations, and Jentic returns their input schema so your agent can trigger and fetch a snapshot. Credentials are injected at call time from your own instance. To run it on your own infrastructure, install Jentic One from its GitHub repo.
Can I restrict what my agent is allowed to do with the Label Studio API?
Yes. Write a rule that allows only the operations your agent needs, such as GET /api/projects/ and POST /api/predictions/, so it can read projects and add pre-labels but cannot delete annotations, and every call it makes is logged by your own instance. You add write or delete operations to the allowed set only when you decide to.
Is there a Label Studio MCP server?
You don't need an MCP server to give your agent Label Studio. Jentic connects it directly from the API Directory: import Label Studio, store your token once, and your agent calls the project, task, and prediction operations on demand without loading another server's tool definitions into its context.
GET STARTED