Reality Kernel
REALITY KERNEL BETA Sovereign Reasoning Engine

Frequently Asked Questions

Everything you need to know about deterministic basin mapping, integration, pricing, and our zero-retention security protocols.

Trial keys are issued by Reality Kernel. Please reach out to our team or contact support through LinkedIn to request a pilot key. Once issued, you can log in directly from the home page.

Reality Kernel maps all system states (files, processes, active connections) into a high-dimensional phase space. Based on mathematical transition rules, we determine if the command's trajectory is attracted to a secure "Basin A" or a dangerous, unstable "Basin B" (reflexive collapse). This is mathematically deterministic, eliminating the hallucinations and latency overhead of probabilistic LLM judges.

A reflexive collapse is a state where an agent's self-generated actions cause the sandbox, operating environment, or parent host to terminate, lock, or compromise itself (e.g., deleting crucial dependency trees or running a reverse-shell). Our governor detects this boundary prior to execution and triggers immediate execution freezes.

The fast path (read-only commands like ls, cat, grep) costs 1 credit. Anything that hits the full engine — risky writes, ambiguous intent, pipelines — costs 5 credits. Every /v1/check response returns the exact cost in credits_consumed.

ALLOW means the action's trajectory is fully secure. WARN means the engine detected suspicious trajectory signals but is not certain — it recommends a human (or your agent's reviewer) makes the override decision. BLOCK means the action's simulation crossed the separatrix into Basin B (irreversible damage, exfiltration, or compromise) and will be frozen immediately.

You can only override a WARN verdict. The premium dashboard exposes Authorize and Terminate buttons when an action is flagged as suspicious. Clicking them logs your override decision directly in the secure audit trail. However, a BLOCK verdict is final and cannot be bypassed or overridden by operators.

Wrap any system or shell action in a call to POST /v1/check with the command and the user's prime intent. If the verdict is ALLOW, run it. If WARN, escalate to a human. If BLOCK, abort. See the Integration Guide for full implementation details.

The fast path returns in under 1 ms. The full engine runs in 50–500 ms — fast enough to wrap any system write operation. Read-heavy agents typically see ~95% fast-path hit rate, so average latency overhead is virtually unnoticeable.

No. Reality Kernel operates on a zero-retention model. Your inputs are evaluated in-memory and hashed. Only the cryptographic proof hashes and basic audit logging properties (credits, timestamps) are persisted in the SOC database. We never train models on your data.

Yes. Old keys are invalidated instantly, and the new key inherits your existing credit balance, company plan, and history logs.

Still have questions? Reach out to Keter Labs — we are ready to assist with any technical or deployment inquiries.
Contact Keter Labs on LinkedIn →