Two sides, code in the center
Focus on the main task on the right and supporting tasks on the left. Each session keeps its context, model and permissions; the editor stays for code.
Explore parallel sessions ↗A practical toolkit for AI development
Take AI beyond chat.
Move development forward.
Connect Claude CLI, model providers and MCP tools inside VS Code. From code and databases to remote hosts and APIs, bring ideas into your daily workflow.
Mix as needed · Bring your own models · Keep your workflow
my-project / Visual Studio Code
PARALLEL SESSION↗ · ↻
Understand data, investigate issues
Inspect the test database’s order tables;
analyze read-only, without changing data.
analyze read-only, without changing data.
Check structure, then relationships.Use configured MCP tools to gather information
and organize findings in a separate context.
and organize findings in a separate context.
↗ mcp · mysql / inspect schema
Continue this task…↑
src / order-service.tsTypeScript
✓ Files, diagnostics and tools in one workspace
MAIN SESSION✎ · ＋ · ↻
Focus on the current task
Implement the order lookup API from the plan;
preserve the existing API behavior.
preserve the existing API behavior.
Start with the existing implementation.Read the code, break down the changes,
then check diagnostics and tests.
then check diagnostics and tests.
✓ read → edit → diagnostics
Describe what you want to build…↑
LLS CCAI · Separate sessions / separate settingsClaude CLI + MCP · Your workspace, your workflow
Built around
your familiar stack
Visual Studio CodeClaude CLIMCPMySQLSQL ServerRedisSSH
Less switching. More building.
One workspace.
Keep the work moving.
Not another chat window: bring conversations, code and tools into the same development workspace.
Tools for real environments
Add SSH, databases, Redis and API debugging through MCP. Install what you need, with this extension or other compatible clients.
See all five tools ↗Your setup, your choices
Choose model providers, inspect usage and manage permissions. Start read-only with limited access and build your workflow gradually.
Read usage guidance ↗Claude Code Config Helper
Your familiar editor.
A more connected AI workflow.
Follow clear setup steps for CLI installation, providers and models, then return your attention to the task at hand.
- Setup guide and CLI download, validation and updates
- Separate models for regular tasks, task flows and compaction
- Seven interface languages, usage and cache statistics
- Workspace MCP configuration and editor diagnostics
From setup to your first task
01
Prepare the CLI
Download or select an existing executable and validate it.
02
Connect a model provider
Add your own provider and available models.
03
Start building
Choose workspace models and a thinking mode, then open chat.
Small tools. Real capabilities.
Five MCP tools. Different environments.
Build more. Switch less.
Start with one extension, add tools as needed, and try them in a test environment first.