Who builds Kerneva
Kerneva is built by Vaibhav Dedhia in Bangalore. It started from a simple observation: every control on an AI agent that moves money checks one transaction at a time, and the agents we measured learned to stay inside that check while their behavior changed underneath it. Kerneva is the control that watches the pattern.
A hosted API and dashboard, a self-hosted image, a Python SDK on PyPI, and a public benchmark across six models. The code has been through three rounds of security audit, one by an outside reviewer.
Direct access to the person who wrote the code. Observation Mode first, so nothing blocks until you have seen what it would have flagged. Self-hosted if your data cannot leave your network.
Working with early teams whose agents issue refunds, payouts or credits. If that is you, twenty minutes on how your agent behaves today is the right first step.
[email protected], or the contact form. Security questionnaires and a DPA are available on request.
How a deployment runs
What changes on Monday morning when Kerneva is installed, and how a team runs it over time.
The operating loop
Connect Kerneva to the agent workflow and begin in Observation Mode.
Kerneva scores actions, records signals, and shows where behavior is drifting.
Operators inspect flagged evaluations and decide whether to allow, escalate, or block.
Thresholds evolve as the team learns what is normal and what is risky.
Every review outcome is recorded so future analysis and audit reflect the full history.
Patterns in history feed back into policy, prompt, and workflow design.
The product is a control loop, not a dashboard.
A dashboard shows data. A control loop closes it: deploy, observe, review, adjust, capture, improve. Every evaluation feeds the next decision.
Customer journey
A realistic adoption path, from first observation to continuous improvement.
What teams do with Kerneva
Review incidents, close cases, and export evidence.
Integrate the SDK, tune boundaries, and connect evaluation outcomes to existing systems.
Validate data flow, isolation, and retention behavior.
Measure whether the team can increase automation without losing control.