Actify

Appearance

Quick Start ​

Get started with Actify in 5 minutes. This guide shows you how to create your first automated workflow using your AI assistant.

Prerequisites ​

Make sure you've completed these steps:

Tutorial: Your First Code Execution ​

Step 1: Hello World ​

Let's verify Actify is working. Simply ask your AI:

Execute this code: console.log('Hello from Actify!')

What happens:

  • Your AI sends the request to Actify
  • Code runs securely in isolation
  • You'll see the output: Hello from Actify!

✅ Success! You've just run your first workflow with Actify.

Step 2: Return Data ​

Now let's get structured data back from your code. Ask:

Run this JavaScript code and show me the result:
const result = {
  message: 'Success',
  timestamp: new Date().toISOString(),
  random: Math.random()
};
return result;

What you'll see: Your AI will receive and display the returned object with your message, timestamp, and a random number.

Step 3: Use NPM Packages ​

Actify automatically installs NPM packages you need. Try this:

Use axios to fetch information about the octocat user from GitHub's API
and show me their name, bio, and public repo count

What happens:

  • Actify automatically installs axios
  • Fetches data from GitHub's API
  • Returns the requested information

Expected result:

json
{
  "name": "The Octocat",
  "bio": "How people build software",
  "public_repos": 8
}

Step 4: Check Your Credits ​

See how many credits you have left:

How many Actify credits do I have?

What you'll see:

  • Current credit balance
  • Credits used this month
  • Usage history

Step 5: View Execution History ​

See all your recent code executions:

Show me my recent Actify executions

What you'll see:

  • List of recent executions
  • Status (completed, failed, running)
  • Timestamp for each execution
  • Credits used per execution

Next: Try These Common Use Cases ​

Data Processing ​

Process this JSON array and return only items with price > 50:
[
  { "name": "Widget A", "price": 25 },
  { "name": "Widget B", "price": 75 },
  { "name": "Widget C", "price": 150 }
]

API Integration ​

Fetch the current weather for New York from wttr.in API
and return just the temperature and conditions

Scheduled Tasks ​

Schedule a job to run every hour that logs the current time
Name it "Hourly Time Logger"

Using Secrets ​

First, store an API key:

Create an Actify secret named GITHUB_TOKEN with value ghp_your_token_here

Then use it in your code:

Execute code that uses my GITHUB_TOKEN secret from the environment
to fetch my GitHub user profile

File Operations ​

Create a JSON file at /tmp/data.json with sample user data,
then read it back and parse it

Understanding What Happens ​

When you ask your AI to execute code with Actify:

  1. Your AI calls Actify - Through the MCP protocol
  2. Code is packaged - Into a secure container with Node.js
  3. Runs in Kubernetes - Isolated sandbox environment
  4. NPM packages installed - If your code uses require() or import
  5. Results returned - Output, return values, and logs sent back
  6. Credits deducted - Based on execution time and resources

Typical execution time: 2-5 seconds for simple code, up to 5 minutes for complex tasks.

Common Patterns & Tips ​

💡 Return Structured Data ​

Always return objects for better AI processing:

Execute code that fetches GitHub stars for nodejs/node repository
and returns: { name, stars, language }

💡 Handle Errors Gracefully ​

Include error handling in your prompts:

Try to fetch data from this API, and if it fails,
return an error object with the error message

💡 Use Descriptive Variable Names ​

This helps your AI understand the results:

Calculate the average price and total quantity from this product data
Return as { averagePrice, totalQuantity, productCount }

💡 Specify Timeouts for Long Tasks ​

For API calls or heavy processing:

Execute this code with a 60 second timeout:
[your code here]

Troubleshooting ​

Code Takes Too Long ​

Symptom: Execution times out

Solution: Ask for a longer timeout:

Execute this with a 120 second timeout: [code]

Package Not Found ​

Symptom: "Cannot find module 'package-name'"

Solution: Verify package name at npmjs.com

Out of Memory ​

Symptom: Process killed, heap error

Solution: Request more memory:

Execute this code with 1GB memory: [code]

Check Execution Logs ​

If something went wrong:

Show me the logs from my last Actify execution

What's Next? ​

Now that you know the basics, explore more:

Learn by Example ​

Understand the Platform ​

Get Help ​

Ready to build real workflows? Check out 40+ Workflow Examples for inspiration!