RasnuOS uses a lightweight, open-source agent that runs as a background service on your server. It initiates secure, outbound WebSocket connections to our control plane. You do not need to open any inbound firewall ports, and it works flawlessly behind NATs or strict corporate firewalls.

No. RasnuOS is entirely stack-agnostic. Whether you run a legacy monolith on a VPS, a complex containerized microservices architecture, or a massive Kubernetes cluster, the AI analyzes your specific environment dynamically and adapts its execution plans accordingly.

The Agent currently supports all major Linux distributions (Ubuntu, Debian, CentOS, RHEL, Alpine, Arch), macOS (Intel and Apple Silicon), and Windows Server (2016+). It auto-detects your architecture during the 10-second installation process.

Yes. RasnuOS is designed to complement existing APM solutions like Datadog, New Relic, or Prometheus. While traditional APMs alert you to problems, RasnuOS actually logs in and resolves them.

The agent is a natively compiled binary that typically consumes less than 15MB of RAM and <1% CPU when idle. It is designed for high-performance environments and only utilizes additional resources when actively streaming logs or executing complex, AI-driven shell plans.

Yes. To fully automate infrastructure tasks like provisioning databases, configuring firewalls, and managing system services, the agent requires root or Administrator privileges. However, every action is strictly governed by the Intent-Driven Execution model.

Yes. All communication between your servers and RasnuOS is end-to-end encrypted via mTLS. We do not store your SSH keys or root passwords. The agent uses secure, ephemeral PTY sessions that are immediately destroyed after a plan finishes executing.

No. RasnuOS uses an Intent-Driven execution model. The AI Architect proposes a Directed Acyclic Graph (DAG) Execution Plan before doing anything. For high-risk operations (e.g., dropping databases, formatting drives, wiping directories), the system halts execution and strictly requires explicit Human-in-the-Loop (HITL) approval via your OS Dashboard.

We utilize a Zero-Knowledge Secrets Vault. Your API keys and environment variables are encrypted at rest and only decrypted and injected directly into the agent's ephemeral session memory at runtime. They never touch your server's physical disk.

We offer granular Role-Based Access Control (RBAC). You can grant junior developers read-only access to view logs and metrics, while restricting execution plan approvals exclusively to senior engineers or administrators.

Every single command, log output, and configuration change is recorded immutably in the Enterprise Ledger and the AI PTY Terminal shard. You have full, irrefutable audit logs for compliance requirements like SOC 2, HIPAA, and GDPR.

The agent is fully stateless. Your server and applications will continue running perfectly without interruption. Once the connection is re-established, the agent will automatically sync the backlog of metrics and resume monitoring.

No. RasnuOS acts as an exoskeleton for your engineering team. It handles the tedious, repetitive toil (like log parsing, scaling, and initial incident triage) so your DevOps team can focus on high-level architecture and feature velocity.

RasnuOS is capable of zero-downtime migrations. The AI provisions the target server, installs exact matching dependencies, copies files securely, and sets up logical database replication. Once data is perfectly synced, it provisions a proxy to reroute traffic while you update your DNS records.

When our Incident Radar detects an anomaly (e.g., CPU spiking, memory leaks, or a dead letter queue filling up), the Swarm immediately wakes up. Agents analyze logs in real-time, trace the stack, and propose an immediate remediation plan (like rolling back a deployment, optimizing a query, or restarting a service) which you can approve with one click.

Absolutely. RasnuOS natively understands container runtimes, orchestration manifests, and all major relational and NoSQL databases. It can execute complex SQL queries, optimize table indexes, or orchestrate rolling container updates completely autonomously.

Yes. Our Vector Memory Shard builds a continuous semantic map of your unique configurations, past incidents, and custom application logic. This means the AutoPilot gets significantly smarter, faster, and more context-aware regarding your specific business logic every single day.

Your base subscription covers platform access, server limits, and user seats. AI operations (like deep server analysis, incident resolution, and automated terminal commands) burn tokens. Tokens are billed precisely by the millisecond from your prepaid wallet. You only pay for exactly what the AI consumes.

Absolutely not. All costs are completely transparent. You set strict limits on your token burn in the Billing Shard, and you will never be charged beyond your prepaid wallet balance.

AI execution will pause gracefully. Your servers will remain online and unaffected, but automated Swarm resolutions, plan generation, and proactive maintenance will halt until you top up your wallet or enable auto-recharge.

Tokens are calculated based on a combination of factors: the complexity of the DAG execution plan, the duration of the active PTY session, and the amount of LLM context window required to analyze your logs and state.

Yes. Enterprise customers receive a 99.99% uptime SLA on the control plane, dedicated account management, and priority 24/7 support for infrastructure emergencies.

Yes. You can upgrade, downgrade, or cancel your subscription at any time directly from the OS billing portal. Plan changes and seat adjustments are pro-rated instantly.