ONEComputer DocsBuild a safe place for AI agents to work.

Install ONEComputer locally, understand how OpenVTC keeps approval separate, and follow the path from an employee workspace to a signed human decision.

Install ONEComputer

See how it works

A computer for each employee

Give people an isolated cloud workspace with the AI agents and tools their work requires.

Clear company rules

Decide what agents can access, what stays blocked, and which actions need a person.

Approval outside the application

Use OpenVTC to keep identity, signing keys, and important decisions in a separate wallet.

Documentation · open source · built for enterprise use

From your first local workspace to a production deployment.

These docs explain how to run ONEComputer, how its safeguards work, and which parts of the complete OpenVTC integration are available today.

01 · Start hereInstall ONEComputer

Use the one-line installer or prepare an existing checkout without losing local state.

02 · BuildRun the local quickstart

Start the database, prepare its tables, and run the portal and policy gateway.

03 · TrustUnderstand OpenVTC approval

Understand why approval happens in an external wallet and how the signed result returns.

04 · OperateDeploy to Azure

See how the services fit together, how releases are checked, and how to monitor their health.

05 · Set rulesDesign clear company policy

Choose what is allowed, blocked, recorded, or sent to a person for review.

06 · Reality checkKnow the current status

See what works today, what is still being integrated, and how completion is verified.

One rule we do not compromise: ONEComputer may show an approval request, but it cannot approve that request itself. A separate OpenVTC wallet signs the manager's decision, and ONEComputer checks that decision before the action continues.