Frequently Asked Questions

Looking for deeper explanations? Explore our guides on login fraud, account takeover, how to detect suspicious logins, and credential stuffing.

Is LoginLlama related to Meta's Llama AI models?

No. LoginLlama is a login security API that scores each login for risk. It is not affiliated with Meta's Llama language models or with any other product that has Llama in its name.

Can I add team members?

Yes! With all paid plans you can add team members.

What is your refund policy?

You can cancel your account via the settings page at any time. Your membership will last until the end of that billing period. If you are in any way dissatisfied, please reach out to me! I would love to hear your feedback. If you are still unhappy, I will refund you personally.

Do you offer free accounts?

Every new account gets 100 free login checks, with no credit card. After that, pick a paid plan. You can upgrade, cancel and pause your subscription at any time.

How do you bill?

Depending on your plan, either monthly or yearly! You can change this at any time. Credits will be added to your account on the day that you are billed.

What payment methods are supported?

Our payment provider is Stripe. We accept all major credit and debit cards. You can input any business tax information on the management page which can be found in your Account Settings.

Are there any limits with the Enterprise plan?

Nope! You can use LoginLlama as much as you want. No limits. We can offer this service because of the serverless nature of the app. It scales according to any and all demands on it. If you would like to run through this alongside your dev team. Please get in touch.

Do you record personal data?

Yes. An IP address and an identity key are personal data, and so is an email address if you send one. You can send an opaque ID as the identity key instead of an email address. Our privacy policy and data processing terms say what we hold and why.

What AI system do you use?

In some cases, if the risk score is high enough, we will use a 3rd party AI system as a final "judge" for the login. This is to prevent false positives. We utilize OpenAI with a variety of models to do this. The "AI Judge" can be disabled in your account settings if you wish.

What data do you record about logins?

LoginLlama records what it needs to rank the risk of a login: the identity key you send, IP address, user agent, browser, operating system, time of day, estimated location from the IP address, and the request source.

Can I self host?

Unfortunately it is not possible to selfhost Loginllama at this time. Please get in touch with us if you would like this service in the future.

I use language X - do you support it?

Perhaps, yes! We have libraries in NodeJS, Python and PHP. Other languages are coming soon. If you would like to request a library, please reach out to us via email - [email protected].

Do you get false positives?

It's understandable to be concerned about false positives. But we have a very low false positive rate. This is because we use a variety of checks (not just AI) to programmatically determine the risk of a login.

How accurate is LoginLlama's suspicious login detection?

LoginLlama achieves approximately 98% accuracy in detecting suspicious logins. The system uses a combination of statistical analysis (Bollinger bands) and AI models to minimize false positives while catching genuine threats. False positive rates are kept low through multi-layered verification, which includes geographic patterns, device fingerprinting, and behavioral analysis.

What programming languages and frameworks are supported?

LoginLlama provides official SDKs for Node.js, PHP, and Python. All SDKs are available on GitHub and npm/Composer/PyPI respectively. If you're using a different language, you can integrate directly via the REST API - it's a simple POST request with JSON payloads. The API documentation includes examples in multiple languages.

How does LoginLlama compare to building in-house?

Building suspicious login detection in-house typically costs $50k-$150k when you factor in engineering time, infrastructure, data storage, and ongoing maintenance. LoginLlama starts at $9 a month. You also avoid the complexity of managing ML models, handling edge cases, and storing historical login data. Most teams can integrate LoginLlama in under an hour versus months of development.

What compliance standards does LoginLlama meet?

We are working towards UK and EU GDPR readiness and have not certified against any standard yet. Today our terms include a data processing agreement, our subprocessors are listed, and your data is stored in the UK. You are the controller and we are the processor, so login checks belong in your own privacy policy. Email [email protected] with a security questionnaire and we will answer it honestly.

Can I customize the sensitivity and detection rules?

Yes. You can adjust the risk threshold for your entire application or configure different sensitivities per user segment. The API returns a numeric risk score (0-10+), so you decide what threshold triggers your security response. You can also disable the AI judge if you prefer purely statistical detection, and configure which signals (location, device, time patterns) carry more weight in the analysis.