Rattled is a Canadian AI research laboratory. We build infrastructure and reference tooling for autonomous AI systems, and we run a security research program that studies how those systems are attacked in the wild.
Autonomous AI agents, Model Context Protocol (MCP) servers, and LLM API surfaces are moving into critical applications faster than the tooling around them is maturing. Rattled is a Canadian AI research laboratory working on that gap — building infrastructure and reference tooling for autonomous systems, and publishing what we learn.
Rattled Intel is our security research program. Rather than operating as a bloated vendor, we run high-interaction honeypot sensors globally to simulate production LLM APIs, function-calling tools, and vector stores — and we publish the resulting telemetry.
By intercepting adversary scans across global cloud networks, we analyze zero-day jailbreak vectors, map tradecraft, and deliver actionable threat intelligence back to developers and security teams.
Realistic trapping across OpenAI, Ollama, LangChain, and MCP tool protocols.
Entropy measurement, prompt injection heuristic classification, and 0-day tagging.
Automated IOC generation for Cloudflare, AWS WAF, CrowdStrike, and Splunk.
Canadian founded, led, and engineered with global cloud sensor coverage.
Have questions about Member access, research partnerships, or security disclosures? Transmit your message directly to our team.