docs: add quickstart guide + Chrome DevTools MCP setup guide

Adds two guide pages referenced by the Chrome DevTools MCP blog post:
- docs/guides/chrome-devtools-mcp-setup.md — browser automation MCP setup
- docs/quickstart.md — 5-minute agent deployment quickstart
- docs/guides/index.md — guide index

Fixes broken links in the Chrome DevTools MCP marketing blog post CTAs:
- /getting-started → /docs/quickstart
- MCP Marketplace → /docs/mcp-server
This commit is contained in:
molecule-ai[bot] 2026-04-21 02:34:41 +00:00 committed by GitHub
parent 6ff0a3bc81
commit 1b84ccac4f
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 182 additions and 0 deletions

View File

@ -0,0 +1,85 @@
---
title: "Chrome DevTools MCP Server Setup"
description: "Connect a Chrome DevTools MCP server to Molecule AI for browser automation in AI agents."
tags: [MCP, browser-automation, guides]
---
# Chrome DevTools MCP Server Setup
Chrome DevTools Protocol (CDP) gives AI agents a real browser — navigate, query, screenshot, and interact with any web page using MCP tool calls.
## Prerequisites
- Google Chrome (desktop) or Chromium
- A Molecule AI workspace
## Start Chrome with Remote Debugging
Chrome exposes CDP over a WebSocket at a configurable port. Start Chrome with remote debugging enabled:
```bash
# macOS
/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome \
--remote-debugging-port=9222 \
--user-data-dir=/tmp/chrome-debug
# Linux
google-chrome --remote-debugging-port=9222 --user-data-dir=/tmp/chrome-debug
# Windows
chrome.exe --remote-debugging-port=9222 --user-data-dir="C:\\tmp\\chrome-debug"
```
To run headless (no visible window):
```bash
google-chrome \
--headless \
--remote-debugging-port=9222 \
--user-data-dir=/tmp/chrome-headless
```
**Note:** Ensure no other Chrome instance is already using port 9222.
## Connect via Workspace Config
In Canvas: **Workspace → Config → MCP Servers → Add browser MCP server**.
Or via API:
```bash
curl -X PATCH https://your-platform.molecule.ai/workspaces/${WORKSPACE_ID}/config \
-H "Authorization: Bearer ${WORKSPACE_TOKEN}" \
-H "Content-Type: application/json" \
-d '{
"mcp_servers": {
"browser": {
"type": "streamable_http",
"url": "http://localhost:9223/mcp"
}
}
}'
```
Replace `http://localhost:9223/mcp` with your MCP server URL — for example, a Cloudflare Workers-deployed CDP bridge, or the Chrome DevTools MCP bridge from the [Chrome DevTools MCP blog post](/blog/chrome-devtools-mcp).
## Available MCP Tools
| Tool | Description |
|------|-------------|
| `browser_navigate` | Navigate to a URL |
| `dom_query` | Query DOM elements via CSS selector |
| `page_screenshot` | Capture a screenshot (PNG) |
| `dom_evaluate` | Execute JavaScript in the page context |
Agents access these tools the same way as any other MCP tool — typed, session-aware, and registered at the workspace level.
## Troubleshooting
**Port 9222 in use:** Find and kill the process with `lsof -i :9222` (macOS/Linux) or `netstat -ano | findstr :9222` (Windows).
**Chrome closes immediately:** Run with `--user-data-dir=/tmp/chrome-debug` to isolate the debugging session from your normal browser profile.
**WebSocket connection refused:** Ensure the MCP server URL is reachable from your Molecule AI deployment.
For full browser automation examples, see the [Chrome DevTools MCP blog post](/blog/chrome-devtools-mcp).

View File

@ -0,0 +1,16 @@
# Guides
Step-by-step guides for common Molecule AI integrations and configurations.
## Browser Automation
- [Chrome DevTools MCP Server Setup](/docs/guides/chrome-devtools-mcp-setup) — Connect a Chrome DevTools MCP server to Molecule AI for browser automation in AI agents.
## Getting Started
- [Quickstart](/docs/quickstart) — Deploy your first AI agent in under 5 minutes.
## Integrations
- [MCP Server Reference](/docs/mcp-server) — Full reference for all supported MCP servers.
- [External Agent Registration](/docs/external-agents) — Register agents from other platforms (Google ADK, OpenCode) in Molecule AI workspaces.

View File

@ -0,0 +1,81 @@
---
title: "Quickstart: Deploy Your First AI Agent"
description: "Get a Molecule AI workspace running in under 5 minutes. Deploy an agent, connect a channel, and send your first task."
tags: [quickstart, getting-started]
---
# Quickstart: Deploy Your First AI Agent
Get a Molecule AI workspace running in under five minutes.
## 1. Install Molecule AI
```bash
git clone https://github.com/Molecule-AI/molecule-core.git
cd molecule-core
docker compose up -d
```
Or use the hosted version at your-platform.molecule.ai.
## 2. Create a Workspace
In Canvas: **Workspaces → New Workspace**. Give it a name and connect your first channel.
Or via API:
```bash
curl -X POST https://your-platform.molecule.ai/workspaces \
-H "Authorization: Bearer ${YOUR_API_KEY}" \
-H "Content-Type: application/json" \
-d '{
"name": "my-first-agent",
"description": "Competitive research assistant"
}'
```
## 3. Configure an MCP Server
For browser automation or custom tool integrations, add an MCP server via Canvas: **Workspace → Config → MCP Servers → Add**.
Or via API:
```bash
curl -X PATCH https://your-platform.molecule.ai/workspaces/${WORKSPACE_ID}/config \
-H "Authorization: Bearer ${WORKSPACE_TOKEN}" \
-H "Content-Type: application/json" \
-d '{
"mcp_servers": {
"browser": {
"type": "streamable_http",
"url": "http://localhost:9223/mcp"
}
}
}'
```
## 4. Connect a Channel
Connect Discord, Telegram, or Slack via the **Channels** tab in Canvas. Paste a webhook URL or bot token — your agent is now accessible from your team's existing communication tools.
## 5. Send Your First Task
```bash
curl -X POST https://your-platform.molecule.ai/workspaces/${WORKSPACE_ID}/tasks \
-H "Authorization: Bearer ${WORKSPACE_TOKEN}" \
-H "Content-Type: application/json" \
-d '{
"task": "Summarize the last 24 hours of deployment events"
}'
```
Or type `/ask what's our deployment status?` in your connected Discord channel.
## What's Next
- [Connect MCP servers for browser automation](/docs/guides/chrome-devtools-mcp-setup)
- [Configure org-scoped API keys for team access](/blog/org-api-keys)
- [Review the full REST API reference](/docs/mcp-server)
- [Set up the agent hierarchy — Community Manager + sub-agents](/docs/architecture)
Explore the [GitHub repo](https://github.com/Molecule-AI/molecule-core) for self-hosting options, or visit [moleculesai.app](https://moleculesai.app) for the hosted platform.