Serverless Code Commands

Serverless code commands let you write and execute JavaScript code directly within Dailybot — no server infrastructure required. Your code runs in Dailybot's secure runtime environment and can make HTTP requests, access persistent storage, and return interactive responses.

Running a Command

Serverless commands can be triggered in three ways:

1. In-Chat Trigger — Users type the command directly in DMs or channels
2. Workflow Trigger — Code runs in response to events or schedules via Dailybot Workflows
3. Command Scheduler — Use the Dailybot SDK to schedule a future run of the command

Command Input

When a command is triggered, your code receives an event object with context about the execution.

Query String

Additional text typed after the command name is available asevent.query. For example, typing /growth_rate July, Augustmakes "July, August" available via event.query.

Context Variables

• event.data.intent — The triggered command name
• event.data.user_full_name — Name of the triggering user
• event.data.targetChannel — Chat channel object where the command executed

// Parse query string input
let query = event.query;
let month = null;
if (query) {
  let parts = query.split(" ");
  if (parts[0]) {
    month = parts[0];
  }
}

return `Generating report for ${month || 'this month'}...`;

Using Serverless Code in Workflows

When the same code runs as a step inside a Dailybot Workflow(Automation), the execution context is different from a chat command. Instead of the chat event fields, your code reads automation data through a workflow object.

• workflow.prev_step — Data returned by the previous step (e.g. workflow.prev_step.result, workflow.prev_step.ai_response)
• workflow.action1 … workflow.actionN — Data returned by a step at a specific position (e.g. workflow.action1.result)
• workflow.context — Metadata about this run (workflow_uuid, workflow_name, organization_uuid, …)

The value you return becomes prev_step.result(and actionN.result) for the steps that follow.

// Read the output of previous steps and pass a value forward
const previous = workflow.prev_step.result;
const aiText = workflow.action1.ai_response;

return `Previous step said: ${previous}`;

Making HTTP Requests

The request variable is available in your code and wraps the superagentlibrary for making HTTP requests.

// GET request
const response = await request.get('https://api.example.com/data');
const data = response.body;

// POST request with JSON body
const response = await request
  .post('https://api.example.com/submit')
  .send({ name: 'Dailybot', type: 'automation' })
  .set('Authorization', 'Bearer your-token');

Returning Results

String Response

Return a simple string to display as the command response:

return "Hello world!";

JSON Response with Interactive Buttons

Return a JSON object with a message and interactive buttons:

return {
  "message": "Last week sales increased by 15% 📈",
  "buttons": [
    {
      "label": "Last month",
      "label_after_click": "Getting sales from last month...",
      "value": "sales last month",
      "button_type": "Command"
    },
    {
      "label": "Last quarter",
      "label_after_click": "Getting quarterly report...",
      "value": "sales last quarter",
      "button_type": "Command"
    }
  ]
};

Dailybot SDK

The DailybotSDK object is available in your code and provides access to persistent key-value storage.

Initialization

DailybotSDK.setAPIKey("your_api_key");

Key-Value Storage

Store JSON objects in user-scoped keys. Data is associated with the user making the request.

// Write data
await DailybotSDK.Storage.write('preferences', {
  theme: 'dark',
  notifications: true
});

// Read data
let prefs = await DailybotSDK.Storage.read('preferences');
// prefs = { theme: 'dark', notifications: true }

// Delete data
await DailybotSDK.Storage.delete('preferences');

Examples

Meeting Room Booking

// Usage: /book-room conference-a 2026-02-15 14:00
const parts = event.query.split(" ");
const roomName = parts[0];
const date = parts[1];
const time = parts[2];

if (!roomName || !date || !time) {
  return "Usage: /book-room <room-name> <date> <time>";
}

// Book the room via external API
const response = await request
  .post('https://rooms.company.com/api/book')
  .send({ room: roomName, date, time });

return `Booked ${roomName} on ${date} at ${time}. Confirmation: ${response.body.id}`;

Tech News Fetcher

// Fetch latest tech headlines
try {
  const response = await request
    .get('https://newsapi.org/v2/top-headlines')
    .query({ category: 'technology', pageSize: 5, apiKey: 'your_api_key' });

  const headlines = response.body.articles
    .map((a, i) => `${i + 1}. ${a.title}`)
    .join('\n');

  return `Today's Tech Headlines:\n${headlines}`;
} catch (error) {
  return "Could not fetch news. Please try again later.";
}

Advanced Patterns

• Use SDK storage to build stateful workflows across multiple command invocations
• Parse event.query to accept subcommands (e.g., /tool help, /tool status)
• Chain commands using interactive buttons for multi-step processes
• Handle errors gracefully and return helpful messages to users
• Use the scheduler to set up recurring data fetches or reports