Skip to content
satiricalguruPublic

About

Forge is a local-first, privacy-respecting AI IDE forked from VS Code. Designed to run completely offline with Ollama, LM Studio, vLLM, and local OpenAI-compatible endpoints—with telemetry completely removed and open extensions via Open VSX.

Topics

Resources

Stars

4 stars

Watchers

0 watching

Forks

Latest commit

 

History

2,810 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Forge Logo

Forge

The Local-First, Privacy-Respecting AI IDE

Forge is an open-source IDE forked from VS Code and built on top of the Void editor codebase. It is designed from the ground up for developers who demand complete code privacy, offline capabilities, and zero telemetry.

License Privacy Telemetry Platform

Key Pillars • Features • Screenshots • Quick Start • Architecture


🧭 Key Pillars

🔒 100% Local & Offline First

Forge completely removes dependency on third-party cloud LLMs. Instead, it provides a built-in provider layer to connect directly with your locally-running models.

🕵️ Zero Telemetry & Privacy Guarding

All tracking, analytics, and metrics-reporting services are completely gutted. Forge will never send your code, prompts, or usage metrics to external servers.

🔌 Open Extensions via Open VSX

Forge uses the open-source Open VSX Registry instead of the proprietary Microsoft Marketplace, giving you access to thousands of extensions without proprietary tracking.


✨ Features

  • Local LLM Providers Built-in: Native support for:
    • Ollama (via /api/chat and /api/generate for FIM/autocomplete)
    • LM Studio (via OpenAI compatible APIs)
    • vLLM, Llama.cpp, and LocalAI out-of-the-box.
  • Provider Auto-Discovery: A background registry service polls your local services (Ollama on 11434, Llama.cpp on 8080, etc.) and displays a live connection status dot:
    • ● Green: Connected and ready.
    • ● Amber: Checking connection status.
    • ● Red: Connection error/offline.
    • ● Grey: Disabled/Unknown.
  • Defence-in-Depth HTTP Guard: Forge's local-provider network utilities accept loopback and explicitly configured private-LAN endpoints, while rejecting public internet destinations. Treat private-LAN providers as trusted because prompts and code are sent to them.
  • Redesigned Agents Window: A dedicated 3-panel workspace interface (Copilot-style) to manage autonomous AI coding agents:
    • Left Sidebar: Session thread browser and Customizations panel with quick emojis (Overview, Agents, Skills, Instructions, Hooks, MCP, Plugins, Tools).
    • Center Panel: The core chat stream, welcome roadmap, auto-approval toggles, and token stream.
    • Right Sidebar: Collapsible view tracking Changes and workspace files side-by-side.
  • React + Tailwind Workbench Integration: Forge compiles React and scopes Tailwind CSS, making it easy to create beautiful, modern UI components directly inside the IDE layout.
  • Code Streaming & Diffing: Inline editing (Cmd+K) and Sidebar Chat (Cmd+L) stream model outputs token-by-token and render side-by-side diffs.

📸 Screenshots

Forge Interface Overview

The elegant, local-first user interface of Forge IDE.


🚀 Quick Start (macOS / Linux)

Prerequisites

  • Node.js version 20.18.2 (configured via .nvmrc)
  • Python (for building native node modules)
  • C++ Compiler (Xcode Command Line Tools on macOS or build-essential on Linux)

Installation & Build

  1. Clone the repository and navigate into it:

    cd Forge
  2. Load the correct Node version and install dependencies:

    source ~/.nvm/nvm.sh && nvm use
    npm install
  3. (Xcode 26+ / Apple Clang 21 on macOS) Patch spdlog build if needed:

    ./scripts/patch-spdlog-xcode26.sh
  4. Build native modules and compile React/Tailwind frontend bundles:

    npm rebuild
    node build/npm/postinstall.js
    npm run buildreact
  5. Compile the main VS Code codebase:

    npm run compile

Running Forge in Developer Mode

To run Forge with live-reloads:

  1. In the first terminal window, start the watcher:

    npm run watch

    (Wait ~2 minutes until compiler checkmarks appear)

  2. In a second terminal window, launch the Forge desktop client:

    ./scripts/code.sh \
      --user-data-dir "$(pwd)/.tmp/user-data" \
      --extensions-dir "$(pwd)/.tmp/extensions"

🛠️ Architecture

Forge extends the VS Code electron-main and renderer architectures:

  • ILocalProvider Interface (src/vs/workbench/contrib/void/common/forgeProviderTypes.ts): Outlines the capabilities of local LLMs including streamChat, streamFIM, healthcheck, and listModels.
  • Auto-Discovery Service (src/vs/workbench/contrib/void/common/forgeProviderRegistryService.ts): Runs a background polling loop to automatically discover active LLM runtimes on localhost.
  • HTTP Guard (src/vs/workbench/contrib/void/common/forge/httpUtil.ts): Intercepts all provider requests and enforces a strict local-only policy, blocking outgoing web requests to remote endpoints.
  • UI Mounts (src/vs/workbench/browser/workbench.ts): Bootstraps React roots to render the sidebar chat, settings page, and Agents window.

📄 License

Forge is licensed under the MIT License and contains source code derived from Microsoft VS Code (MIT) and Void (Apache 2.0). See LICENSE.txt for details.

About

Forge is a local-first, privacy-respecting AI IDE forked from VS Code. Designed to run completely offline with Ollama, LM Studio, vLLM, and local OpenAI-compatible endpoints—with telemetry completely removed and open extensions via Open VSX.

Topics

Resources

Stars

4 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages