Identity tells you who an agent is. It doesn't decide what the agent is allowed to do. If your agents take real actions — call APIs, push infra changes, send messages, move money — Kynara is the authorization, approval, and audit layer that sits in front. Join a small group of teams shaping it, free.
Full access through the program and 12 months free afterward, at whatever tier you need. No card, no procurement — a design-partner letter, not an MSA.
A shared Slack channel and biweekly working sessions. Your top requests get prioritized on the roadmap.
We pair with you to wire Kynara into your stack — SDK or the MCP Gateway — and maintain the integration for your framework.
Tamper-evident decision logs, policy export, and a SOC 2 / EU AI Act Article 12 evidence pack you can put in front of auditors.
The program is built for teams running agents that take consequential actions and have someone who owns that risk. Two profiles fit best: DevOps / SecOps teams whose agents execute remediation (agentic SOC, AI SRE, incident response), and regulated teams (fintech, healthcare, insurance) running internal agents under audit pressure.
We pick one real agent workflow and the three policies that matter most for it.
We pair with you to gate that workflow via the SDK or MCP Gateway — first in a non-prod environment, then live.
Over 60–90 days we expand scope, tune policies and approvals, and capture a concrete outcome — something you can point to.
Wrap up with verified audit, an optional case study, and your free year. Logo and quote are always opt-in.
Tell us a bit about your agents. We read every application and reply within two business days.
Prefer to talk first? Email [email protected] or book a scoping call.
Yes — free through the program and for 12 months after. We're optimizing for feedback and reference outcomes, not revenue, at this stage.
No. A logo, quote, or case study is opt-in and only if Kynara delivered real value. You can participate fully and stay anonymous.
A 45-minute scoping call, one workflow integrated (~30 days, we pair with you), and a 30-minute biweekly call. That's it.
Identity platforms authenticate the agent — who it is. Kynara is the layer above: it decides allow / deny / require-approval on every action, guarantees an agent can't exceed the human who dispatched it, and keeps a tamper-evident log. See how Kynara compares.