Agents run where the work already is.

The model does not need your laptop. The work does — the repo, the credentials, the local context. Station is the part of Pardx that lives on your machines and runs agents there, under your rules.

Connections → Connect a machine

Connect this machine

Download Pardx Station for your OS, install it, then enter the one-time pairing code from Connections → Connect a machine. The code works once, expires, and is linked to your account.

Early access rolls out in small groups — when your spot opens, Station and pairing are ready in Connections.

Run work on connected machines.

Station lets agents operate where the repo, credentials, and local context already live. Laptop, dev server, cloud box — pair any of them, and dispatch from any browser.

Approve actions before they cross a boundary.

Risky writes, pushes, and scoped actions pause for explicit Approval. Allow once, allow for the run, or stop everything — from wherever you are, including your phone.

Keep one record across engines.

Each run leaves context behind for the next one: what ran, what changed, and what was approved. The Brain carries it forward, whichever engine did the work.

  • Pair from Connections: download Station, then enter the one-time pairing code.
  • Detects what you already have: Claude Code, Codex, Cursor, Antigravity, git, node, python, docker.
  • Dispatches Claude Code, Codex, Cursor, and Antigravity under your policy.
  • A local policy engine decides what may run; everything else waits for you.
  • Every action is written to a local audit log with its time and reason.
  • The desktop app is signed and notarized for macOS. Windows and Linux are next.

Run agents without giving up control.

Request early access