← Documentation

PLUXIX GUIDE

Getting started

Install the Windows Agent, sign in, choose a local workspace and run your first safe task.

1. Install PLUXIX Agent

Download the latest approved Windows build from the PLUXIX download page. Use the installer for a persistent installation or the portable build for a temporary test. The Agent includes its local runtime and keeps its service boundary private to Windows.

  • Use the newest stable build unless your Admin has asked you to test another channel.
  • Allow the Windows service prompt when installing; the Agent does not need a Windows account password for normal setup.
  • Do not expose the local runtime port to the public internet.

2. Sign in

Sign in with your PLUXIX account using email and password, Google or GitHub when those methods are enabled. The browser and the installed Agent use the same account, but the account session and local workspace remain separate from provider credentials.

  • If two-step verification is enabled, enter the code sent to your account email.
  • If a session expires, sign in again; do not paste tokens into a task.
  • Use the public domain https://pluxix.com for account and connection flows.

3. Choose a workspace

A workspace is the local folder the Agent is allowed to inspect and change. Select the repository root when possible so project files, Git history and tests are visible together.

  • Use a separate folder or branch for risky experiments.
  • Confirm the folder before allowing file or terminal actions.
  • Keep credentials, .env files and generated secrets outside prompts and task text.

4. Start with a useful brief

Describe the outcome, constraints and acceptance criteria. Include the command that should pass or the observable behavior that proves the work is complete.

  • Good: fix the failing checkout test, preserve the API response shape and run the relevant test suite.
  • Avoid: make everything better.
  • Ask the Agent to inspect first when you do not yet know the affected files.