Azure Machine Learning Web Service API Key

Detects Azure Machine Learning web service API key patterns. This pattern is based on a Microsoft Purview built-in sensitive information type. Users already running Purview may prefer to enable the built-in SIT directly, or use this version as a starting point for customisation.

Type

regex

Engine

universal

Confidence

high

Confidence justification

High confidence: structurally constrained Base64 pattern with Machine Learning context keywords reduces false positive rates significantly. Added context gating and exclusion rules improve precision and reduce incidental matches.

Detection quality

Verified

Jurisdictions

global

Regulations

Criminal Code Act 1995 (Cth)

Frameworks

CIS Controls, ISO 27001, NIST CSF, PCI-DSS, SOC 2

Data categories

credentials, security

Scope

specific

Risk rating

10

Platform compatibility

Purview: Compatible, GCP DLP: Compatible, Macie: Compatible, Zscaler: Compatible, Palo Alto: Degraded, Netskope: Unsupported

Pattern

(?i)(?:ml[._-]?api[._-]?key|scoring[._-]?key|web[._-]?service[._-]?key|azureml[._-]?key)\s*[:=]\s*"?[A-Za-z0-9+/]{30,60}={0,2}"?

Corroborative evidence keywords

Machine Learning, Azure ML, AzureML, web service, scoring endpoint, api key, ml studio, workspace, api_key, apikey, access key, access token, auth token, authorization, bearer, conn str, connection string, connectionstring, cookie, credential (+44 more)

Proximity: 300 characters

Should match

• ml_api_key=AAAA0000BBBB1111CCCC2222DDDD3333EEEE4444FFFF5555 — Azure ML web service API key
• scoring_key: "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghij==" — ML scoring endpoint key
• azureml_key="00000000000000000000000000000000000000000000" — Azure ML key with placeholder value

Should not match

• ml_api_key="" — Empty key value
• workspaceName=my-ml-workspace — Workspace name, not a key
• template example placeholder record identifier — Template/sample context should be excluded even when anchor words are present

Known false positives

• Azure ML documentation and quickstart guides with placeholder API keys. Mitigation: Check for common placeholder patterns and documentation context.
• API keys from other services in similar Base64 format. Mitigation: Require proximity to Machine Learning specific keywords for differentiation.

References

• Azure Machine Learning web service API key entity definition - Microsoft Learn