Invoking Automations

Making API Calls

Once you have an endpoint and API key, you can trigger automations from anywhere.

Basic Request

curl -X POST "https://api.getgranite.ai/api/your-org/your-endpoint" \
  -H "X-Granite-API-Key: gk_live_abc123..." \
  -H "Content-Type: application/json" \
  -d '{"param1": "value1"}'

Request Format

Headers

Header	Required	Description
`X-Granite-API-Key`	Yes	Your API key
`Content-Type`	Yes (for POST)	`application/json`

Body

JSON object with your parameters:

{
  "invoiceNumber": "INV-2024-001",
  "amount": 1500.00,
  "vendor": "Acme Supplies"
}

Response Modes

Asynchronous (Default)

Request returns immediately:

{
  "runId": "run_abc123",
  "status": "queued",
  "message": "Automation queued successfully"
}

Then poll for status:

curl "https://api.getgranite.ai/api/agent-runs/run_abc123" \
  -H "X-Granite-API-Key: gk_live_abc123..."

Synchronous

Add ?wait=true to wait for completion:

curl -X POST "https://api.getgranite.ai/api/your-org/endpoint?wait=true" \
  -H "X-Granite-API-Key: gk_live_abc123..."

Response:

{
  "runId": "run_abc123",
  "status": "completed",
  "duration": 45000,
  "result": {
    "output": "Report generated successfully",
    "filePath": "reports/january-2024.pdf"
  }
}

Synchronous calls have a timeout (default 5 minutes). For long automations, use async.

Status Values

Status	Meaning
`queued`	Waiting for a driver
`running`	Currently executing
`completed`	Finished successfully
`failed`	Encountered an error
`cancelled`	Stopped manually

Error Handling

HTTP Errors

Code	Meaning	Action
`400`	Bad request	Check parameters
`401`	Unauthorized	Check API key
`404`	Not found	Check endpoint URL
`429`	Rate limited	Wait and retry
`500`	Server error	Contact support

Execution Errors

{
  "runId": "run_abc123",
  "status": "failed",
  "error": {
    "type": "element_not_found",
    "message": "Could not locate the Submit button",
    "screenshot": "https://..."
  }
}

Code Examples

JavaScript
Python
TypeScript SDK

async function runAutomation(params) {
  const response = await fetch(
    'https://api.getgranite.ai/api/my-org/my-endpoint',
    {
      method: 'POST',
      headers: {
        'X-Granite-API-Key': process.env.GRANITE_API_KEY,
        'Content-Type': 'application/json',
      },
      body: JSON.stringify(params),
    }
  );

  if (!response.ok) {
    throw new Error(`HTTP ${response.status}`);
  }

  return response.json();
}

// Usage
const result = await runAutomation({
  month: 'January',
  year: 2024,
});
console.log('Run ID:', result.runId);

import requests
import os

def run_automation(params):
    response = requests.post(
        'https://api.getgranite.ai/api/my-org/my-endpoint',
        headers={
            'X-Granite-API-Key': os.environ['GRANITE_API_KEY'],
            'Content-Type': 'application/json',
        },
        json=params
    )
    response.raise_for_status()
    return response.json()

# Usage
result = run_automation({
    'month': 'January',
    'year': 2024
})
print(f"Run ID: {result['runId']}")

import { client } from '@getgraniteai/ts-sdk';

client.setConfig({
  headers: {
    'X-Granite-API-Key': process.env.GRANITE_API_KEY,
  },
});

const result = await client.invokeOrgEndpoint({
  orgIdentifier: 'my-org',
  slug: 'my-endpoint',
  body: {
    month: 'January',
    year: 2024,
  },
});

Polling for Status

For async calls, poll until completion:

async function waitForCompletion(runId) {
  while (true) {
    const response = await fetch(
      `https://api.getgranite.ai/api/agent-runs/${runId}`,
      {
        headers: { 'X-Granite-API-Key': apiKey },
      }
    );
    const data = await response.json();

    if (['completed', 'failed', 'cancelled'].includes(data.status)) {
      return data;
    }

    await new Promise(r => setTimeout(r, 2000)); // Wait 2 seconds
  }
}

Webhooks

For event-driven workflows, you can set up a webhook to be called when execution completes. (Coming soon)

Get Started

Getting Started

Dashboard

Agent Execution

RPA Automation

Public API

VM Management

Driver

Troubleshooting

Invoking Automations

Making API Calls

Basic Request

Request Format

Headers

Body

Response Modes

Asynchronous (Default)

Synchronous

Status Values

Error Handling

HTTP Errors

Execution Errors

Code Examples

Polling for Status

Webhooks

Next Steps

API Reference

TypeScript SDK

​Making API Calls

​Basic Request

​Request Format

​Headers

​Body

​Response Modes

​Asynchronous (Default)

​Synchronous

​Status Values

​Error Handling

​HTTP Errors

​Execution Errors

​Code Examples

​Polling for Status

​Webhooks

​Next Steps

API Reference

TypeScript SDK

Making API Calls

Basic Request

Request Format

Headers

Body

Response Modes

Asynchronous (Default)

Synchronous

Status Values

Error Handling

HTTP Errors

Execution Errors

Code Examples

Polling for Status

Webhooks

Next Steps