AI Cyber Assistant

What the Cyber Assistant can help with

The Cyber Assistant is a small language model that runs on our own server through Ollama. It is tuned for defensive work around the tools on this site: reading DNS and email-authentication results, explaining TLS and certificate errors, suggesting hardening steps, and turning a goal into a concrete command for Linux, macOS or Windows.

Example questions

• "My DMARC report shows SPF pass but DMARC fail. Why?"
• "Write an nginx config that enables HSTS and a strict Content-Security-Policy."
• "Explain what X25519MLKEM768 means in a TLS handshake."
• "Give me a dig command to check whether my zone is signed with DNSSEC."
• "What does 'unable to get local issuer certificate' mean and how do I fix it?"

Privacy

Prompts are processed locally and are not sent to a third-party AI provider. We do not store conversations or log their content. Even so, never paste passwords, private keys, API tokens or personal data into the chat. Redact log lines before sharing them.

Limitations

This is a compact model, so it can be wrong, outdated or overconfident. Treat answers as a starting point and verify them with the live tools (for example the SPF checker, SSL checker or PQC scanner) and with the official RFCs and vendor documentation. The assistant refuses requests aimed at attacking systems you do not own.

Frequently asked questions

Is the assistant free?

Yes. There is a short per-minute limit per visitor to keep the shared model responsive.

Why is it offline sometimes?

The model runs on a single server. During maintenance or heavy load the status indicator turns red. Try again a few minutes later.

Can it run a scan for me?

No. It explains and suggests, but it does not query domains itself. Run the relevant tool, then paste the result here and ask what it means.