Meet the future of IDEs: the Integrated Agents Environment.
Homepage | Agents Extension | Agents MCP
Text editors became IDEs. Now IDEs become IAEs. Swarmify turns your editor into a command center for orchestrating Claude, Codex, Gemini, and Cursor in parallel — with full visibility into every agent's work.
Install the Agents Extension from VS Code/Cursor marketplace, then add Agents MCP:
# Claude Code
claude mcp add --scope user Swarm -- npx -y @swarmify/agents-mcp
# Codex
codex mcp add swarm -- npx -y @swarmify/agents-mcp@latest
# Gemini CLI
gemini mcp add Swarm -- npx -y @swarmify/agents-mcpRun your first swarm:
/swarm Ship billing polish - 70% Claude planning, 30% Codex coding
| Step | What Happens |
|---|---|
| 1 | Run /swarm "your task with mix needs" |
| 2 | Lead agent creates a plan and assembles the team |
| 3 | You review and approve; agents execute in parallel |
When you're running 10+ orchestrators each spinning 10+ agents, terminals and TUIs collapse. You need an environment built for that scale:
| Agent | CLI | Best For |
|---|---|---|
| Claude | claude |
Planning, synthesis, multi-step reasoning |
| Codex | codex |
Fast implementation, surgical edits |
| Gemini | gemini |
Research depth, multi-system changes |
| Cursor | cursor-agent |
Debugging, tracing through codebases |
Different models have different strengths. Swarmify lets you use them all — Claude plans while Codex implements while Cursor debugs, in parallel.
CLI agents are powerful — but running them in separate terminal windows means constant context-switching. You lose sight of your code, your agents, and what each one is doing.
The Agents Extension gives you:
- Full-screen terminals - Agents run as editor tabs, not buried panels
- Session persistence - Close VS Code, reopen, your agent tabs come back
- Sub-agent spawning - Any agent can delegate to other agents via Agents MCP
- Approval gates - You approve plans before agents execute
- Keyboard-first - 12+ shortcuts for spawning, switching, labeling agents
Your IDE becomes an IAE. Each agent runs as a full-screen editor tab. Each agent can spawn sub-agents. You orchestrate.
| Shortcut | Action |
|---|---|
Cmd+Shift+A |
Spawn new agent |
Cmd+Shift+L |
Label agent by task |
Cmd+Shift+D |
Open Dashboard |
Cmd+Shift+G |
Generate git commit |
Cmd+Shift+H/V |
Split horizontal/vertical |
MCP server that enables sub-agent spawning. Any agent can spawn any other agent — Claude can spawn Codex, Gemini can spawn Claude.
4 tools: Spawn, Status, Stop, Tasks
| Mode | File Access | Use Case |
|---|---|---|
plan |
Read-only | Research, code review |
edit |
Read + Write | Implementation, refactoring |
ralph |
Full + auto-loop | Autonomous via RALPH.md |
Slash commands calibrated for each agent. Commands prefixed with s spawn multiple agents:
| Command | Description |
|---|---|
/swarm |
Multi-agent orchestration |
/plan / /splan |
Design approach / parallel validation |
/debug / /sdebug |
Root cause / parallel investigation |
/test / /stest |
Critical paths / parallel testing |
/ship / /sship |
Pre-launch / parallel verification |
Multi-agent debugging:
/sdebug The API returns 500 errors intermittently on /users endpoint
Feature with mixed agents:
/swarm Add dark mode - 50% Claude planning, 30% Codex coding, 20% Cursor testing
Code review:
/sconfirm Review the last 3 commits for security issues
Swarmify is free and open source.
Each agent uses your own API keys. Spawning 3 Claude sub-agents means 3x your normal Claude API cost. No hidden fees.
- VS Code or Cursor
- At least one agent CLI (
claude,codex,gemini, orcursor-agent) - API keys configured for your agents
The Agents CLI has moved to its own repository: @swarmify/agents-cli
MIT