1. Install and open Elcro

Install from Releases, then launch Elcro.

2. Sign in (for Agent / Tab / Ctrl+K)

Open Settings → Account and complete desktop SSO. The editor works offline for editing; AI surfaces need an account.

3. Add a model provider

1. Open Settings → Models.
2. Fill API keys for OpenAI, Anthropic, Gemini, OpenRouter, etc., or point Ollama at http://127.0.0.1:11434.
3. Refresh / enable the models you want for Chat.

4. Open a folder

File → Open Folder so tools can see your workspace. Many agent tools require a workspace root.

5. Chat

1. Press Ctrl/Cmd+L (or open the Elcro sidebar).
2. Set mode to Agent.
3. Pick a strong model (e.g. Claude Sonnet, GPT-4.1, or a capable local coder).
4. Ask Elcro to inspect a file and make a small edit.

6. Optional: MCP

Create ~/.elcro/mcp.json:

{
  "mcpServers": {
    "playwright": {
      "command": "npx",
      "args": ["-y", "@playwright/mcp@latest"]
    }
  }
}

Then Settings → Tools & MCPs → Reload, toggle the server on, and ask Agent to use MCP on a URL.

Related

• First AI Chat
• Model Configuration