Pylon FAQs

Question 1

Is Pylon easy to integrate and secure for local development?

Accepted Answer

Yes, Pylon is designed for simplicity with zero dependencies, no proxy, and a single HTTP endpoint (`http://127.0.0.1:27681/mcp`) for AI agent configuration. For security, all servers bind exclusively to `127.0.0.1`, making it safe for local development without network exposure.

Question 2

How does Pylon handle multiple VS Code windows?

Accepted Answer

Pylon features a hub/satellite architecture. Each VS Code window runs an AgentServer (satellite), while the first window opened establishes a central hub. AI agents can then target specific VS Code windows by their unique session IDs for concurrent project management.

Question 3

What is Pylon and what does it do?

Accepted Answer

Pylon is a VS Code extension that exposes various IDE functionalities like terminals, diagnostics, debugging, and editor actions to AI coding agents (e.g., Claude Code, Gemini) through a local HTTP endpoint, enabling AI to interact with your active VS Code environment.

Question 4

What kind of VS Code features can AI agents access through Pylon?

Accepted Answer

AI agents can access a rich set of tools including terminal commands (run, read output), diagnostics (errors, warnings), IDE functions (get symbols, find references, rename symbols, run VS Code commands), editor actions (open files, format, organize imports), and comprehensive debugging controls (breakpoints, start/stop debug, inspect state).

Question 5

Can I control which VS Code features AI agents can use?

Accepted Answer

Absolutely. Pylon offers granular tool access control, allowing you to enable or disable specific tools (e.g., `debug_start`, `rename_symbol`) via a visual picker in the VS Code Command Palette or by manually configuring the `pylon.disabledTools` setting. Changes take effect immediately.

Pylon

Pylon

주요 기능

사용 사례

주요 기능

사용 사례