Skip to content
| Marketplace
Sign in
Visual Studio Code>Programming Languages>MiraBridge AINew to Visual Studio Code? Get it now.
MiraBridge AI

MiraBridge AI

MiraBridge AI

|
55 installs
| (1) | Free
Control AI coding agents from VS Code with Desktop, mobile approvals, terminal visibility, file review, and current Claude, GPT, and Gemini models.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Copied to clipboard
More Info

MiraBridge AI

MiraBridge AI

AI coding control for VS Code, Desktop, and mobile
Run agent workflows in your local workspace, then review and approve them from any MiraBridge surface.

VS Code Marketplace Website Desktop app License


What is MiraBridge?

MiraBridge AI gives developers a practical control layer for AI-assisted coding sessions across VS Code, the MiraBridge Desktop app, and mobile approval surfaces. Start work in your local workspace, guide the agent, review code changes, watch terminal output, and keep approvals organized without tying one runner to another.

Use it when an AI coding workflow needs more than a single prompt: planning, file edits, terminal commands, build feedback, review, and human approval.

  • Agent workflows in VS Code: chat, plan, build, fix, debug, and review from one sidebar.
  • Desktop-native runner path: use MiraBridge Desktop as an independent app for agent work outside VS Code.
  • Mobile approval loop: pair your phone, monitor progress, and approve or reject tool actions remotely.
  • Terminal and workspace visibility: follow command output, diagnostics, file context, and build status.
  • Multi-provider model routing: work with supported Claude, GPT, Gemini, and GLM models through one interface.
  • Developer-owned control: keep review, approval, credentials, and workspace context under your direction.

Core Workflows

AI Chat and Agent Mode

Ask questions, explain code, generate changes, or move into agent mode for multi-step work. MiraBridge keeps the conversation connected to your workspace, files, diagnostics, and terminal context.

Code Review and Tool Approval

Review file operations and tool actions before they land. Approve, reject, or redirect the workflow so the agent can help without taking control away from you.

Terminal and Build Monitoring

Run commands, inspect output, track failures, and keep build or test feedback visible. MiraBridge is built for the loop where the next useful action depends on what the terminal just said.

Desktop and Mobile Surfaces

Use the VS Code extension and MiraBridge Desktop as independent runners over the same backend Gateway and agent services. Pair the mobile app for approval and monitoring without requiring Desktop and VS Code to depend on each other.

Mobile Pairing

Pair the VS Code extension with the MiraBridge mobile app using QR code pairing. Continue monitoring active work, approve agent actions, and respond to prompts when you step away from the keyboard.

Model Selection

Use supported AI providers from a single workflow:

Provider Use case
Anthropic Claude Claude Sonnet 4.6 and Claude Opus 4.7 for coding, reasoning, refactoring, and long context
OpenAI GPT GPT-5.5, GPT-5.2, and GPT-5.1 Codex for implementation, review, and coding workflows
Google Gemini Gemini 3.1 Flash and Gemini 2.5 Pro for fast responses and broad assistant workflows
GLM GLM-5 for additional managed model routing

Bring your own API keys where supported by your MiraBridge setup.


Quick Start

1. Install

Search for MiraBridge AI in the VS Code Extensions panel, or run:

ext install MiraBridge.mirabridge-vscode-extension

2. Sign In

  1. Open the Command Palette with Cmd+Shift+P on macOS or Ctrl+Shift+P on Windows/Linux.
  2. Run MiraBridge: Login.
  3. Sign in to your MiraBridge account.

3. Start a Session

Open the MiraBridge activity bar panel, choose a model, and send a request. You can also pair the mobile app from MiraBridge: Pair with Mobile App.


Commands

Command Description
MiraBridge: Login Sign in to your account
MiraBridge: Open AI Chat Open the MiraBridge chat panel
MiraBridge: New Conversation Start a fresh conversation
MiraBridge: Select Model Choose the active AI model
MiraBridge: Start Agent Start an agent workflow
MiraBridge: Auto-Approve Permissions Configure tool approval behavior
MiraBridge: Pair with Mobile App Connect your phone by QR code
MiraBridge: Open Settings Review extension settings

Keyboard Shortcuts

Shortcut Action
Cmd+Shift+M Open MiraBridge Chat
Cmd+Shift+N New Conversation
Cmd+Shift+E Explain Selected Code
Cmd+Shift+R Refactor Selected Code
Cmd+Shift+F Fix Errors
Cmd+Shift+T Toggle Chat/Agent Mode

Configuration

Search for MiraBridge in VS Code settings.

Setting Description
mirabridge.ai.provider Preferred AI provider
mirabridge.ai.model Preferred model
mirabridge.gateway.url MiraBridge Gateway URL
mirabridge.autoApprove.enabled Enable or disable automatic tool approval

Legacy mirabridge.ai.defaultProvider and mirabridge.ai.defaultModel settings are migrated automatically.


Security

  • Credentials are stored with VS Code SecretStorage.
  • Workspace Trust is respected for file and command operations.
  • Gateway communication uses secure WebSocket connections.
  • Tool execution is visible and can require explicit approval.
  • Mobile pairing is scoped to the connected workspace session.

Requirements

  • VS Code 1.85.0 or later
  • Internet connection for MiraBridge Gateway and AI provider access
  • MiraBridge account

Links

  • Website
  • Desktop App
  • Documentation
  • GitHub
  • Issue Tracker
  • Mobile App

License

MIT License. See LICENSE for details.

  • Contact us
  • Jobs
  • Privacy
  • Manage cookies
  • Terms of use
  • Trademarks
© 2026 Microsoft