burp-ai-agent

Why we included this project

Anyone who does pentesting or bug bounty work inside Burp Suite probably knows the drill: copy a request out of Burp, paste it into a chat window, paste the answer back. This extension removes that step by wiring an LLM directly into the tool. It supports eleven AI backends, from local models like Ollama and LM Studio to cloud and CLI providers, and exposes dozens of MCP tools so an external agent can drive Burp on its own. The passive and active scanners use the model to surface likely vulnerabilities while you stay in manual testing, and the privacy handling holds up: stored API keys are encrypted, request and response bodies can be redacted with custom regex patterns, and a pre-send tripwire warns before high-entropy secrets leave the tool. If you want AI assistance without giving up control over what leaves the tool, this is worth a look.