Hosted MCP Quick Reference

Quick reference for using ERA Agent’s hosted/cloud MCP server with Claude Desktop.

Setup (One-Time)

{
  "mcpServers": {
    "era-agent": {
      "url": "https://anewera.dev/mcp/v1"
    }
  }
}

Config Locations:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Linux: ~/.config/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

Then: Restart Claude Desktop (Cmd+Q, not just close)

Quick Tool Reference

Language Tools (Instant Execution)

Tool	Use For	Example
`era_python`	Python scripts	FizzBuzz, data analysis
`era_node`	JavaScript/Node.js	Array operations, algorithms
`era_typescript`	TypeScript code	Type-safe execution
`era_deno`	Deno runtime	Modern JS/TS with Deno APIs
`era_shell`	System commands	`pip install`, `npm install`

Session Tools (Stateful Work)

Tool	Use For	Example
`era_create_session`	Start persistent environment	Create Python session
`era_run_in_session`	Run code in session	Execute with state
`era_list_sessions`	View all sessions	Check active sessions
`era_get_session`	Get session details	Check session status
`era_update_session`	Update session config	Change timeout, network access
`era_delete_session`	Clean up session	Remove unused session

File Operations (Session Only)

Tool	Use For	Example
`era_upload_file`	Add files to session	Upload CSV, config
`era_read_file`	Read session files	View results
`era_list_files`	List session files	Check workspace

Common Patterns

Quick Python Script

Run this Python code:
print("Hello World")
print([x**2 for x in range(10)])

Install Packages & Run

1. Create a Python session
2. In that session, run: pip install pandas requests
3. Then run your code using those packages

Web Scraping (Network Included!)

Create a Python session, install requests and beautifulsoup4,
then scrape the top stories from Hacker News

Multi-Language Testing

Test this sorting algorithm in both Python and JavaScript

Data Pipeline

1. Create a Python session
2. Upload data.csv with [your data]
3. Install pandas
4. Analyze the data and save results
5. Read the results file

Long-Running Task (Update Timeout)

1. Create a session for ML training
2. Update the session to extend timeout to 10 minutes
3. Run training code that takes 8 minutes
4. Check results

Key Differences: Hosted vs Local

Feature	Hosted (Cloud)	Local (Self-Hosted)
Setup	Just add URL ✅	Need Docker/Firecracker
Network	Always enabled ✅	Disabled by default
Config	URL only	Binary path + env vars
State	Cloud (Durable Objects)	Local filesystem
Scaling	Automatic ✅	Manual
Maintenance	Zero ✅	You manage it
Internet Access	Built-in ✅	Must enable per session

Gotchas

✅ Works Great

F-strings: print(f'{x}: {y}')
Multi-line code with indentation
Special characters in strings
Complex nested structures
Internet access (enabled by default!)

⚠️ Watch Out

No State Between Tools: Use sessions for stateful work
Packages Not Installed: Use era_shell to install first
File Paths: Use relative paths in sessions, upload files first

Default Limits

Resource	Default	Notes
CPU	1 core	Per execution
Memory	256 MiB	Configurable
Timeout	30 sec	Configurable
Network	Enabled	Always available
Storage	Cloud (R2)	Persistent across requests

Troubleshooting

Problem	Solution
Tools not showing	Restart Claude Desktop (Cmd+Q)
Execution fails	Check package dependencies
Package not found	Use `era_shell` to install
File not found	Upload file first with `era_upload_file`
Network error	Should work automatically (enabled by default)

Best Practices

When to Use What

Use Language Tools (era_python, etc.):

Quick one-off scripts
Simple calculations
Testing snippets
Independent executions

Use Sessions:

Installing packages
Multi-step workflows
Stateful computations
File processing pipelines

Use era_shell:

Package installation (pip, npm)
System commands (ls, mkdir)
Environment setup
NOT for running code (use language tools)

Performance Tips

Network Included: No need to enable - perfect for web scraping!
Cloud State: Sessions persist across requests
Reuse Sessions: Create once, use multiple times
Install Once: Install packages at session start

Example Prompts

Data Science

Create a Python session, install pandas and matplotlib,
analyze this CSV data, create a visualization, and save it

API Integration

Use era_python to fetch data from this API, process it,
and display the results (network access included!)

Algorithm Development

Implement binary search in Python with logging,
test it with various inputs, then implement in JavaScript and compare

Web Development

Create a Node.js session, install express,
create a simple REST API, and test it

Cloud Benefits

✅ Zero Infrastructure: No Docker or Firecracker to manage ✅ Global Edge: Cloudflare’s network for low latency ✅ Auto Scaling: Handles load automatically ✅ Persistent State: Sessions survive across requests ✅ Cloud Storage: Files stored in R2 ✅ Always Online: No local agent to start/stop ✅ Network Enabled: Internet access built-in