# Meet Mini Hound.

> A smaller, less powerful version of BloodHound. Paste a website URL or add a small code file. Mini Hound explains possible security problems and shows what it read.

Official source: https://warhoundai.com/mini-hound/
Updated: 2026-11-03

- Available now: Free and Plus plans
- Access: Verified email and website required
- Token allowance: $0.50 in monthly credits; each window starts with its first model request
- Per message: Up to 3 public resources; one website at a time

## Start with a website or a code file

Create an account, verify your email and sign in. Activate Free or Plus through checkout, then verify your website by publishing your account’s code at /verify/bloodhound, accept the acceptable use policy, then paste a URL on that same verified website. You can also attach a small text or code file.

Verification codes expire after 24 hours; ownership verification lasts 30 days. Other websites and subdomains are blocked. Mini Hound reads public pages and linked code when needed. It points to the source behind a possible issue, explains why it matters and suggests a fix. You can ask follow-up questions in the same chat.

- [Open Mini Hound](https://warhoundai.com/mini-hound/app/)
- [Create an account](https://warhoundai.com/sign-up/)

## What it can help with

Mini Hound v1 (Gemini 3.8 Flash) and Mini Hound v1 Pro (Gemini 3.1 Pro Preview) through Google Cloud is the current model and provider.

A possible problem is not the same as a proven vulnerability. A real finding needs evidence about how the application behaves and what it is meant to allow. Mini Hound should say when that evidence is missing.

- Explain a small piece of code in plain language.
- Look for visible patterns that may lead to problems, such as unsafe HTML rendering.
- Fetch public HTTPS pages and linked source within the three-resource limit.
- Read and search collected files, consult the private 428-entry reference and the preserved legacy checklist, and save source-backed candidate findings.
- Load cloud conversations and target/finding history across devices.
- Help you understand a reported issue and what to check next.

## What it cannot do

Conversations are saved in the cloud under your account and sync across devices. Older browser chats are imported only when you choose to import them. Messages, attached text and collected source are sent to the AI provider for processing. Do not send passwords, API keys or personal data.

- Fetch private backend code or access authenticated accounts from a public URL. Small source excerpts you supply can still be reviewed.
- Run penetration tests, browser exploits, high-volume requests or denial-of-service tests.
- Review every file on a large website in one message.
- Guarantee that a website is safe or that it has found every issue.

- [Acceptable use policy](https://warhoundai.com/acceptable-use/)
- [How we handle data](https://warhoundai.com/privacy/)
- [Read the test results and limits](https://warhoundai.com/mini-hound/benchmarks/)

## Mini Hound and BloodHound

Mini Hound is the smaller public product you can use now. It uses Mini Hound v1 (Gemini 3.8 Flash) and Mini Hound v1 Pro (Gemini 3.1 Pro Preview) through Google Cloud and bounded tools for public HTTPS collection, stored-source reading/search, checklist lookup and source-backed candidate recording. The Free plan has $0.50 in monthly credits, 2 messages per minute, and shared capacity limits. Requests can be delayed or fail when a provider is unavailable.

BloodHound is our larger security product in development. Broader assessments, continuous monitoring, editor plugins and browser integrations remain planned. The private MCP and customer review API are available to accounts with an active server-granted API entitlement.

- [Learn about the larger BloodHound product](https://warhoundai.com/bloodhound/)
- [Product availability](https://warhoundai.com/product-status/)
- [Pricing](https://warhoundai.com/pricing/)