Skip to main content

Tensorfleet VS Code Extension

Build, simulate, and operate PX4/ROS 2 robots without leaving VS Code. This guide gets you from install to a project with dashboards and telemetry streaming.

Install​

  1. Prereqs: VS Code (latest), Node.js 20+ for building from source, and Git.
  2. Get the extension:
    • Marketplace: search for "TensorFleet Robots and Drone Environment".
    • Local build: clone vscode-tensorfleet, run ./build.sh, then code --install-extension dist/tensorfleet-*.vsix.
  3. Reload VS Code so the Tensorfleet activity bar icon appears.

Create a workspace​

  1. Open the folder where you want to work.
  2. Press Cmd/Ctrl+Shift+P → TensorFleet: Install Bundled Tools to download the curated ROS 2, Gazebo, and mission assets.
  3. Start a project:
    • Bottom TensorFleet Tools panel → 🚀 New Project, or
    • Command palette → TensorFleet: Create New Project.
  4. The scaffold includes config/drone_config.yaml, example missions, and launch files so the status bars and dashboards light up automatically.

Operate from the dashboards​

  • Open all panels: TensorFleet: Open All Dashboards (or open individual Gazebo, QGroundControl, AI Ops, and ROS 2 panels from the sidebar).
  • Connect to ROS 2/PX4: TensorFleet: Connect to ROS2 and TensorFleet: Start PX4 Telemetry Monitor.
  • Visualize & teleop: TensorFleet: Open Image Panel (Option 3 - React) and TensorFleet: Open Teleops Panel (Option 3 - React).
  • Project-aware status bars: the ROS version and drone status items appear when a Tensorfleet project marker (e.g., config/drone_config.yaml or src/main.py) is present.

AI + MCP (optional)​

  • Start the MCP server so AI assistants can drive panels and tools: TensorFleet: Start MCP Server.
  • View configuration to point Cursor/Claude at the server: TensorFleet: Show MCP Configuration.
  • Stop when you are done: TensorFleet: Stop MCP Server.

Troubleshooting​

  • Check the TensorFleet output channel for messages like TensorFleet MCP Bridge started.
  • If dashboards do not appear, verify you are inside a Tensorfleet project (markers mentioned above) and reload the window.
  • For ROS 2 connectivity, make sure the ROS environment is sourced before launching VS Code, or restart after sourcing.