SDK: Python

Suspicious login detection for Python

Score logins in Python apps with a lightweight API call.

Install command

pip install loginllama==3.1.0

Add LoginLlama to Python in one call

Call LoginLlama from your Python login handler before you create the session. Send the email, IP, and user agent, then allow, challenge, or block based on the returned risk score.

Python login check
from loginllama import LoginLlama, AuthenticationOutcome
from flask import Flask, request

app = Flask(__name__)
loginllama = LoginLlama(api_token=os.environ.get('LOGINLLAMA_API_KEY'))

# Use middleware for automatic IP/User-Agent detection
@app.before_request
def setup_loginllama():
    loginllama.middleware()()

# Pre + Post Auth Pattern
@app.route('/login', methods=['POST'])
def login():
    email = request.form['email']
    password = request.form['password']

    try:
        # Step 1: Pre-auth check
        pre_check = loginllama.check(
            email,
            authentication_outcome='pending'
        )

        if pre_check.risk_score > 7:
            return {'error': 'Login blocked'}, 403

        # Step 2: Authenticate
        auth_result = authenticate(email, password)

        # Step 3: Report outcome
        if auth_result.success:
            loginllama.report_success(auth_result.user.id)
            return {'success': True}
        else:
            loginllama.report_failure(email)
            return {'error': 'Invalid credentials'}, 401
    except Exception as error:
        print(f'LoginLlama error: {error}')
        return {'success': True}  # Fail open on errors

Why teams choose LoginLlama for Python

Score every login in under 100ms

Detect VPNs, proxies, and TOR exit nodes

Explain why a login is risky with clear risk reasons

Catch errors from the check and logins go ahead during an outage

How it works

Step 1

Install or connect LoginLlama

Choose the SDK or integration point that fits your auth flow and add your API key.

Step 2

Check risk at login time

Send the email, IP, and device context before completing authentication.

Step 3

Act on the risk score

Allow, challenge, or block the login based on risk score and reasons.

Signals we analyze

New device or browser fingerprints

Impossible travel and velocity checks

Bot-like login bursts and credential stuffing

Geo anomalies and risky IP reputation

Works well with your Python stack

DjangoFlaskFastAPICelery workers

Python login security FAQ

How do I add suspicious login detection to Python?

Install LoginLlama with `pip install loginllama==3.1.0`, then call the risk check from your Python login handler before you create the session. Send the user's email, IP address, and user agent, and branch on the returned risk score.

Will LoginLlama slow down Python logins?

No. Risk scoring returns in under 100ms. If LoginLlama is unreachable the check throws an error, so catch it and let the Python login go ahead. An outage then never blocks a real user.

What does LoginLlama detect on Python logins?

New devices and browser fingerprints, impossible travel and login velocity, credential-stuffing bursts, and traffic from VPNs, proxies, TOR exit nodes, and high-risk IPs — each returned with a clear, human-readable reason code.

Does LoginLlama work with Django, Flask, FastAPI, Celery workers?

Yes. LoginLlama runs at the login step itself, so it works alongside Django, Flask, FastAPI, Celery workers and the rest of your Python stack without replacing your existing authentication provider.

Other integrations

LoginLlama works with the same single API call across every stack.

Python login protection

Protect Python users from suspicious logins

Add LoginLlama in minutes and block risky logins in real time.

Start with 100 free checks