Skip to content
| Marketplace
Sign in
Visual Studio Code>Other>AIOT CodeNew to Visual Studio Code? Get it now.
AIOT Code

AIOT Code

株式会社AIoT

|
5 installs
| (0) | Free
AI development workbench with Lark SSO and remote agent backend.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Copied to clipboard
More Info

aiot-code

aiot-code is a VS Code extension scaffold for a development-focused AI agent, powered by a local local_dev_agent HTTP server. It provides:

  • A dedicated Activity Bar entry and sidebar panel
  • Active workspace / file / selection context collection
  • Direct integration with a local development agent runtime
  • Patch suggestions with diff preview before apply
  • Command suggestions with explicit confirmation before execution

Prerequisites

Make sure Ollama is installed and running (used by the local agent internally):

ollama serve
ollama pull gemma4:latest

The local_dev_agent package must be available at the configured agent path (default: /Users/zy/DOG):

# Verify the package can start the HTTP server
cd /Users/zy/DOG
PYTHONPATH=/Users/zy/DOG python3 -m local_dev_agent.server --host 127.0.0.1 --port 8420

Quick Start

  1. Install dependencies:

    npm install
    
  2. Build the extension:

    npm run compile
    
  3. Press F5 in VS Code to launch an Extension Development Host.

  4. The extension will automatically check http://127.0.0.1:8420/health. If the agent is not running, it will try to start python3 -m local_dev_agent.server --host 127.0.0.1 --port 8420 for you. Configure in VS Code settings:

    • aiotCode.agentPath — Parent directory containing the local_dev_agent package (default: /Users/zy/DOG)
    • aiotCode.pythonPath — Python interpreter path (default: python3)
    • aiotCode.model — Ollama model name passed to the agent (default: gemma4:latest)

How It Works

When you send a message in the sidebar panel:

  1. The extension builds a prompt including your message, active file content, and selection context.
  2. It checks whether the local HTTP agent is healthy on 127.0.0.1:8420.
  3. If needed, it starts python3 -m local_dev_agent.server with PYTHONPATH=<agentPath>.
  4. The extension sends the request to POST /run or POST /run/stream.
  5. The agent processes the prompt internally (may use tools like file read, grep, shell) and returns a final text answer over HTTP/SSE.
  6. The extension displays the reply in the conversation panel.

Notes

  • No external gateway or API key is required — the agent runs locally.
  • If aiotCode.agentPath points at the package directory itself, startup will fail. It must point to the parent directory that contains local_dev_agent.
  • The agent has its own tool runtime (file, grep, shell) and can delegate to subagents (plan, explore, verify).
  • The MVP intentionally avoids automatic file edits and silent command execution on the VS Code side.
  • Contact us
  • Jobs
  • Privacy
  • Manage cookies
  • Terms of use
  • Trademarks
© 2026 Microsoft