Does Ragna give an AI unrestricted administrator access?
No. Authorized users, machines, roles and tool policies define what is allowed. Sensitive operations can require explicit approval.
RAGNA · PRODUCT OVERVIEW
Ragna gives development teams an independent execution path for AI coding, plus a permission-controlled Agent for working on real Windows, macOS and Linux machines. Use a Workspace, machine permissions, OAuth and a credit-backed Wallet; Ragna does not bypass the quotas of other AI providers.
Ragna is a software platform for AI coding continuity and secure connections between MCP-compatible AI clients and authorized machines. Choose a Ragna Cloud execution path or connect your own machine through an outbound Ragna Agent. Workspace identity, RBAC, tool policy and approvals constrain what AI can do.

| Criteria | Ragna Cloud | Connected machine |
|---|---|---|
| Where work runs | Managed cloud execution | Your enrolled Windows, macOS or Linux machine |
| Setup | Start from your Workspace | Install/enroll the outbound Agent and grant access |
| Machine files and Git | Cloud environment scope | Only machine resources allowed by policy |
| Billing | Ragna Wallet and plan credit | Ragna Wallet and plan credit; check live terms |
| Approvals and audit | Workspace policy | Workspace, machine policy and audit |
Ragna combines a Workspace control plane, an independent execution path for AI coding and an MCP connection layer. An AI client can request a structured tool; Ragna checks identity and policy before invoking an authorized tool. MCP describes tools and resources, but is not a replacement for server-side access control or a guarantee that every client supports every integration.
Enroll your own machine with the Ragna Agent, select its Workspace and grant the smallest necessary permission. Connect Ragna in a compatible ChatGPT environment through the supported OAuth flow, then begin with a read-only request such as listing connected machines or inspecting a service. Never paste machine passwords or SSH keys into chat.
No. External providers control their own quotas, usage windows and policies. Use the official reset, credit or plan options for those services. Ragna offers a separate execution path, billed by Ragna credit, that may support continuity for your workflow; it does not modify third-party limits.
An outbound Agent connects an authorized machine to a Workspace. Ragna evaluates identity, roles, machine permissions and tool restrictions; sensitive operations may require explicit approval. Plans cover a credit baseline and top-ups extend credit for heavier work. The live pricing page is the source of current prices and available options.
No. Authorized users, machines, roles and tool policies define what is allowed. Sensitive operations can require explicit approval.
No. The machine Agent connects outbound in the normal Ragna architecture. A newly published SSH, RDP or MCP port is not needed for regular use.
Ragna is an independent platform. It can connect compatible AI clients via supported MCP/OAuth flows but cannot change provider quotas or third-party plan requirements.
Ragna does not promise unlimited free use. Plans and Wallet credits apply. Check the live pricing page for current packages and usage terms.