Questions, answered.
Forward deployed security engineers, the Security Factory, and AI Traffic Control.
AKA embeds forward deployed security engineers with AI-first teams. They build the controls the environment needs in the Security Factory and keep them current.
A senior security engineer who works inside your team, in your channels, repositories and review queue. They build the controls your agents need and ship them to production.
Flightcrew, six agents that write and test each control, and AI Traffic Control, which enforces them. Engineers run both on every engagement.
Agents run through four layers: orchestration, tool calls, identity and data access. Together that is the agent harness. AI-TC checks every prompt and tool call across all four before it runs.
AKA's open-source security engine for AI agents. It checks every prompt and tool call and keeps regulated data on your machine.
Yes. AI-TC and the AKA kits are on GitHub under open-source licenses.
Community is free for individual use. Enterprise adds one signed policy across the organization and a record of what every endpoint enforces.
AI-TC supports Claude Code and Claude Desktop today. claude-tools and preflight-skills run on Claude Code, Codex and Antigravity.
Detection and enforcement run locally, so PCI, PHI and PII never reach a model or a third party. The one exception is optional setup calibration. If you turn it on, AI-TC sends what its first history scan finds to the model provider to be rated. It never goes to AKA, and you can withdraw consent at any time.
Yes. AKA is SOC 2 Type II and ISO 27001 certified. Both reports are available on request from legal@akasecurity.io.
Talk to an engineer, or try AI-TC for free on GitHub.