CAI v1.0: The Cybersecurity CLI Agent, Refined

Alias Robotics -

From prototype to production: CAI reaches v1.0

After months of iteration, community feedback and relentless testing in real-world engagements, we are proud to announce CAI v1.0, the definitive release of our cybersecurity AI agent platform.

This is not an incremental update. CAI v1.0 represents a fundamental leap in usability, reliability and operational effectiveness for cybersecurity professionals.

What is new in CAI v1.0

Improved MCP and Burp Suite compatibility. Integration with the tools you already use has been completely reworked. MCP server connections are more stable and Burp Suite workflows are now seamless, whether you are running automated scans or manual testing through the proxy.

More agile operation. CAI v1.0 is faster and more responsive. Context switching between tasks, tool invocations and multi-step reasoning chains have all been optimized for the pace of real engagements.

Better behavior in longer sessions. Extended penetration testing sessions demand consistency. We have addressed context degradation, improving how CAI maintains coherence and task awareness across long, complex operations. The agent stays sharp from the first reconnaissance command to the final report.

Fewer refusals. We have pushed our already industry-leading permissiveness even further. CAI v1.0 delivers fewer interruptions and more actionable output across the full spectrum of cybersecurity tasks. From vulnerability discovery to exploitation and post-exploitation.

Automated agent selection. No more manually choosing the right agent for the job. CAI v1.0 automatically selects the optimal agent configuration based on your task context. Start a web assessment and it configures for web. Pivot to binary analysis and it adapts. Simpler operation, better results.

Prompt suggestions for day-zero users. New to CAI? The platform now offers intelligent prompt suggestions that guide you through common workflows from the moment you launch. No documentation required to start being productive. Whether you are a seasoned red teamer or just exploring AI-assisted security, CAI v1.0 meets you where you are.

Better experience, same mission

Our mission is to make cybersecurity radically more accessible. CAI exists to empower our users with the best specialized models in the industry. Every feature in v1.0 was designed with one question in mind: does this help a cybersecurity professional do their job better? If the answer was not a clear yes, it did not make the cut.

We believe the community should use whatever tools work best for them. As we do not plan to compete with scaffolding providers, we offer support for Claude Code, Codex and OpenCode.

Get started with CAI v1.0

CAI v1.0 is available now. Upgrade to the latest version and join the change. The new prompt suggestions will guide you through your first session, no prior experience needed.

We built CAI for the cybersecurity community. v1.0 is our commitment to that mission, delivered.