OpenClaude VS Code

The open AI coding assistant for VS Code — powered by any LLM you choose.

OpenAI · Anthropic · Google Gemini · DeepSeek · Ollama · AWS Bedrock · Vertex AI · GitHub Models · 200+ OpenAI-compatible endpoints.

Why OpenClaude?

Modern AI coding assistants are powerful — but they lock you into one provider, one billing account, one pricing tier. OpenClaude flips the model: one polished VS Code UI, any LLM backend you want.

Already paying for OpenAI GPT-4o? Use it.
Prefer Claude Sonnet via Anthropic's direct API? Use it.
Running Ollama or LM Studio locally for privacy? Use it.
On an enterprise plan with AWS Bedrock or Vertex AI? Use it.
Hitting rate limits? Switch providers mid-session with /provider.

OpenClaude is a full-featured VS Code extension that wraps the open-source OpenClaude CLI. The CLI is where all the intelligence lives — tool use, provider routing, MCP, slash commands. The extension gives you a first-class editor experience on top: streaming chat panel, native diff viewer, @-mentions, session history, checkpoints, and more.

Features

Chat & Conversation

Streaming chat panel with markdown rendering and syntax-highlighted code
Tool-call visualization — collapsible blocks show what the AI is reading, editing, and running
Session history — browse, resume, or fork any past conversation
Checkpoint / rewind — snapshot the workspace and restore to any point
Stop / interrupt generation at any time; no half-finished edits

Native VS Code Integration

Diff viewer — AI-proposed changes open in VS Code's built-in diff editor with Accept / Reject buttons in the editor title bar
@-mentions — reference files, folders, symbols, and line ranges for precise context
Status bar with live token count and cost
Git worktree support — run parallel AI sessions on the same repo without conflicts
Onboarding walkthrough for first-time users

Multi-Provider Support

Switch between LLM providers on the fly via /provider, the provider badge, or env vars:

Provider	Models	Setup
OpenAI	GPT-4o, GPT-4 Turbo, GPT-4o-mini	`OPENAI_API_KEY`
Anthropic	Claude Opus, Sonnet, Haiku	Claude Code OAuth or `ANTHROPIC_API_KEY`
Google Gemini	Gemini 2.0 Flash, Pro	`GOOGLE_API_KEY`
Ollama	Llama 3, Mistral, CodeLlama (local, free)	`OPENAI_BASE_URL=http://localhost:11434/v1`
DeepSeek	DeepSeek V3, R1	OpenAI-compatible endpoint
AWS Bedrock	Claude via Bedrock	AWS credentials
Google Vertex AI	Claude via Vertex	GCP credentials
GitHub Models	Various via GitHub Marketplace	GitHub PAT
Codex (ChatGPT)	gpt-5.4, codexplan, codexspark	`OPENAI_BASE_URL=https://api.codex.openai.com/v1`
Custom	Any OpenAI-compatible endpoint	`OPENAI_BASE_URL`

Developer Tools

5 permission modes — Default, Plan, Accept Edits, Bypass, Don't Ask
MCP (Model Context Protocol) server integration — extend the AI with your own tools
Plugin manager — install / update / manage MCP plugins from inside the editor
Slash commands — /commit, /review, /diff, /resume, /compact, /mcp, and more
Environment variable injection per workspace
Respects .gitignore in file searches by default

Install

From the VS Code Marketplace (recommended)

Search for OpenClaude in the Extensions panel (Ctrl+Shift+X / Cmd+Shift+X) and click Install, or:

code --install-extension HarshAgarwal1012.openclaude-vscode

From a `.vsix` file (latest dev build)

Download the latest .vsix from Releases, then:

code --install-extension openclaude-vscode-0.2.5.vsix

Prerequisites

OpenClaude requires the underlying CLI:

npm install -g @gitlawb/openclaude

(The extension is a thin UI wrapper — all AI intelligence lives in the CLI.)

Quick Start

1. Install the CLI

npm install -g @gitlawb/openclaude

2. Configure a provider

OpenAI (simplest):

export CLAUDE_CODE_USE_OPENAI=1
export OPENAI_API_KEY=sk-your-key-here
export OPENAI_MODEL=gpt-4o

Anthropic (native Claude):

export ANTHROPIC_API_KEY=sk-ant-your-key
# Claude Code OAuth also works if you're already signed in

Google Gemini:

export CLAUDE_CODE_USE_GEMINI=1
export GOOGLE_API_KEY=AIza-your-key
export GEMINI_MODEL=gemini-2.0-flash

Ollama (local, free, private):

ollama serve  # start Ollama first
export CLAUDE_CODE_USE_OPENAI=1
export OPENAI_API_KEY=ollama
export OPENAI_BASE_URL=http://localhost:11434/v1
export OPENAI_MODEL=llama3

AWS Bedrock:

export CLAUDE_CODE_USE_BEDROCK=1
export AWS_REGION=us-east-1
# Standard AWS credentials chain (env / ~/.aws / IAM role)

Any OpenAI-compatible endpoint (DeepSeek, Together, Fireworks, OpenRouter, …):

export CLAUDE_CODE_USE_OPENAI=1
export OPENAI_API_KEY=sk-your-key
export OPENAI_BASE_URL=https://api.deepseek.com/v1
export OPENAI_MODEL=deepseek-chat

Or just open OpenClaude and use /provider to set up providers interactively.

3. Open OpenClaude

Press Cmd+Escape (macOS) / Ctrl+Escape (Windows/Linux), or
Click the OpenClaude icon in the Activity Bar, or
Run OpenClaude: Open in New Tab from the Command Palette

4. Start coding

Type your prompt. Use @ to mention files, / for slash commands. The AI edits stream into a VS Code diff view — accept, reject, or let it keep going.

Keyboard Shortcuts

Action	macOS	Windows / Linux
Open / Focus OpenClaude	`Cmd+Escape`	`Ctrl+Escape`
Open in new tab	`Cmd+Shift+Escape`	`Ctrl+Shift+Escape`
Insert @-mention	`Alt+K`	`Alt+K`
New conversation	`Cmd+N`	`Ctrl+N` (opt-in, see settings)

Slash Commands

Type / in the chat input to browse all available commands. Highlights:

Command	Description
`/provider`	Set up and switch LLM providers
`/model`	Switch between models for the current provider
`/compact`	Compact conversation context to save tokens
`/resume`	Browse and resume past sessions
`/diff`	Show current git diff in the chat
`/commit`	Ask the AI to create a git commit
`/review`	Review code, a diff, or a PR
`/mcp`	Manage MCP servers
`/plugins`	Manage OpenClaude plugins
`/help`	Show all commands

Settings

All settings live under openclaudeCode.* in VS Code settings:

Setting	Type	Default	Description
`openclaudeCode.selectedModel`	string	`"default"`	AI model to use
`openclaudeCode.initialPermissionMode`	enum	`"default"`	Starting permission mode
`openclaudeCode.useCtrlEnterToSend`	boolean	`false`	Require Ctrl+Enter to send (vs plain Enter)
`openclaudeCode.preferredLocation`	enum	`"panel"`	Default panel location
`openclaudeCode.autosave`	boolean	`true`	Auto-save before AI reads or writes
`openclaudeCode.respectGitIgnore`	boolean	`true`	Honor `.gitignore` in file searches
`openclaudeCode.useTerminal`	boolean	`false`	Launch in terminal mode instead of webview
`openclaudeCode.environmentVariables`	array	`[]`	Extra env vars passed to the AI process
`openclaudeCode.hideOnboarding`	boolean	`false`	Hide the onboarding checklist
`openclaudeCode.enableNewConversationShortcut`	boolean	`false`	Enable Cmd/Ctrl+N to start a new conversation

Architecture

┌─────────────────────────────────┐
│  Webview (React + Tailwind)     │  ← UI: chat panel, diff, mentions
└───────────────┬─────────────────┘
                │ postMessage
┌───────────────▼─────────────────┐
│  Extension Host (TypeScript)    │  ← VS Code integration, permissions, sessions
└───────────────┬─────────────────┘
                │ stdin / stdout NDJSON
┌───────────────▼─────────────────┐
│  OpenClaude CLI (child process) │  ← Intelligence: tools, providers, MCP, plugins
└───────────────┬─────────────────┘
                │ OpenAI Chat Completions API
┌───────────────▼─────────────────┐
│  Any LLM provider               │  ← OpenAI / Anthropic / Gemini / Ollama / …
└─────────────────────────────────┘

The extension is deliberately thin. All provider logic, tool execution, MCP server plumbing, and slash-command handling happens inside the CLI — so upgrading the brain means npm install -g @gitlawb/openclaude@latest with no VS Code reinstall needed.

Contributing

git clone https://github.com/Harsh1210/openclaude-vscode
cd openclaude-vscode
npm install
cd webview && npm install && cd ..
npm run build

Development (watch mode):

npm run watch
# Press F5 in VS Code to launch an Extension Development Host

Run tests:

npm test

Package a .vsix:

npx @vscode/vsce package --no-dependencies --allow-missing-repository

Issues and PRs welcome — see the issue tracker.

License

MIT — see LICENSE.

Keywords: vscode extension · ai coding assistant · claude code alternative · openai gpt-4o · claude sonnet · gemini · ollama · aws bedrock · local llm · mcp · model context protocol · diff viewer · chat panel · typescript

Name		Name	Last commit message	Last commit date
Latest commit History 69 Commits
.github/workflows		.github/workflows
.vscode		.vscode
docs		docs
resources		resources
src		src
test		test
webview		webview
.eslintrc.json		.eslintrc.json
.gitignore		.gitignore
.openclaude-profile.json		.openclaude-profile.json
.prettierrc		.prettierrc
.vscodeignore		.vscodeignore
CHANGELOG.md		CHANGELOG.md
LICENSE		LICENSE
README.md		README.md
esbuild.config.mjs		esbuild.config.mjs
openclaude-settings.schema.json		openclaude-settings.schema.json
package-lock.json		package-lock.json
package.json		package.json
tsconfig.json		tsconfig.json
vitest.config.ts		vitest.config.ts

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

OpenClaude VS Code

Why OpenClaude?

Table of Contents

Features

Chat & Conversation

Native VS Code Integration

Multi-Provider Support

Developer Tools

Install

From the VS Code Marketplace (recommended)

From a `.vsix` file (latest dev build)

Prerequisites

Quick Start

1. Install the CLI

2. Configure a provider

3. Open OpenClaude

4. Start coding

Keyboard Shortcuts

Slash Commands

Settings

Architecture

Contributing

License

About

Uh oh!

Releases 1

Packages

Uh oh!

Contributors

Uh oh!

Languages

Folders and files

Latest commit

History

Repository files navigation

OpenClaude VS Code

Why OpenClaude?

Table of Contents

Features

Chat & Conversation

Native VS Code Integration

Multi-Provider Support

Developer Tools

Install

From the VS Code Marketplace (recommended)

From a .vsix file (latest dev build)

Prerequisites

Quick Start

1. Install the CLI

2. Configure a provider

3. Open OpenClaude

4. Start coding

Keyboard Shortcuts

Slash Commands

Settings

Architecture

Contributing

License

About

Topics

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases 1

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

From a `.vsix` file (latest dev build)

Packages