Issue a token and use MLflow from a notebook

The platform includes MLflow, so with a single token, experiments you run in your own notebook (Jupyter and the like) can also be recorded on the platform's Experiments (실험) screen. The token carries the tenant it was issued in, and records go into that tenant only.

Go to Manage MLflow tokens (MLflow 토큰 관리) under account settings in the account menu at the top right (or Admin console (관리자 콘솔) → MLflow tokens (MLflow 토큰)). Set ① a name (for example my-notebook) and ② the number of days until expiry (default 90 days; No expiry (무기한) is also possible), and ③ press Issue (발급).
MLflow tokens — ① name ② days until expiry ③ Issue
The token is shown only this once. Copy the two lines below with Copy command (명령 복사) and keep them somewhere safe.
```
export MLFLOW_TRACKING_URI=https://mlops.example.com/mlflow
export MLFLOW_TRACKING_TOKEN=<your-secret>
```

Start the notebook from the same shell, or set the environment variables in the notebook's first cell, and use MLflow as usual.

import os, mlflow
os.environ["MLFLOW_TRACKING_URI"] = "https://mlops.example.com/mlflow"
os.environ["MLFLOW_TRACKING_TOKEN"] = "<your-secret>"

mlflow.set_experiment("helmet-notebook")
with mlflow.start_run(run_name="baseline"):
    mlflow.log_param("imgsz", 640)
    mlflow.log_metric("mAP50", 0.81)

The helmet-notebook experiment and its run appear on the platform's Experiments (실험) screen. You can also put a finished run into the registry with Register model on the run detail.

The MLflow client sends MLFLOW_TRACKING_TOKEN as an Authorization: Bearer header. The token determines the tenant, so you do not need to state it separately.
See the Active (활성), Expired (만료됨) and Revoked (회수됨) status in the token list, and Revoke (회수) tokens you no longer use. Tenant admins see every token in the tenant.
Treat tokens like passwords. Do not write them into notebook files and share them; keep them in environment variables or a secret store.

For detailed integration such as model logging rules, how to register in the registry and container registry tokens, see the Integration chapter.