Webhook Integration: Receive Unpod Call Data in Your System

Last updated: July 22, 2026 · API version: v2 · Tested against: openapi.yaml 2.0.0

Every Unpod call produces rich structured data: full transcripts, AI-generated summaries, sentiment analysis, call recordings, and extracted lead intelligence. This guide shows you how to pull that data into your own system - whether that’s a CRM, database, Slack, or a custom pipeline.

What Data Is Available After a Call

When an Unpod call completes, the task output contains:

{
  "output": {
    "call_id": "675272289779974401",
    "call_type": "outbound",
    "call_status": "completed",
    "start_time": "2026-01-31 11:47:38.067021",
    "end_time": "2026-01-31 11:48:11.385745",
    "duration": 33,
    "recording_url": "https://storage.example.com/call.ogg",
    "transcript": [
      { "role": "assistant", "content": "Hi, this is Aria from Acme...", "timestamp": "..." },
      { "role": "user", "content": "Yes, I'm interested", "timestamp": "..." }
    ],
    "post_call_data": {
      "summary": {
        "status": "Connected",
        "summary": "Prospect expressed interest in the enterprise plan...",
        "callback_datetime": "2026-02-01T10:00:00"
      },
      "profile_summary": {
        "tone": "Positive",
        "engagement": "High",
        "interest_level": "Very Interested",
        "outcome": "Meeting Scheduled",
        "next_action": "Send proposal email"
      },
      "classification": {
        "labels": ["Interested", "Decision Maker"]
      }
    }
  }
}

This is the data you want flowing into your business systems automatically.

Integration Architecture

Unpod uses a poll + process model. You poll the API to detect completed calls, then process and forward the data.

Animated post-call data integration diagram showing Unpod API data flowing through a polling service into CRM, Slack, database, and custom webhook systems.

For high-volume deployments, run the polling service as a background worker. For lower volume, a cron job works fine.

Step 1: Set Up Your Webhook Receiver

Build an HTTP endpoint that receives processed call data. Here’s a minimal receiver in Node.js and Python:

const express = require('express');
const app = express();
app.use(express.json());

app.post('/webhook/unpod-call', (req, res) => {
  const { task_id, run_id, status, output, input } = req.body;

  if (status !== 'completed') {
    return res.sendStatus(200); // ignore non-completed
  }

  const {
    call_status,
    duration,
    recording_url,
    transcript,
    post_call_data
  } = output;

  const summary = post_call_data?.summary;
  const profile = post_call_data?.profile_summary;

  console.log(`Call completed: ${task_id}`);
  console.log(`Contact: ${input.name} (${input.contact_number})`);
  console.log(`Duration: ${duration}s | Outcome: ${profile?.outcome}`);
  console.log(`Next action: ${profile?.next_action}`);

  // Process further: update CRM, send Slack, write to DB...
  processCallData({ task_id, input, summary, profile, recording_url, transcript });

  res.sendStatus(200);
});

app.listen(3000, () => console.log('Webhook receiver running on :3000'));

from fastapi import FastAPI
from pydantic import BaseModel
from typing import Optional, Any

app = FastAPI()

class CallWebhookPayload(BaseModel):
    task_id: str
    run_id: str
    status: str
    input: dict
    output: Optional[dict] = None

@app.post("/webhook/unpod-call")
async def receive_call(payload: CallWebhookPayload):
    if payload.status != "completed":
        return {"ok": True}

    output = payload.output or {}
    post_call = output.get("post_call_data", {})
    profile = post_call.get("profile_summary", {})
    summary = post_call.get("summary", {})

    print(f"Call completed: {payload.task_id}")
    print(f"Contact: {payload.input.get('name')} ({payload.input.get('contact_number')})")
    print(f"Outcome: {profile.get('outcome')}")
    print(f"Next action: {profile.get('next_action')}")

    # Process further
    await process_call_data(payload.task_id, payload.input, output, post_call)

    return {"ok": True}

Step 2: Build the Polling Service

This service runs continuously, detects newly completed calls, and forwards data to your webhook receiver.

const axios = require('axios');

const UNPOD_API = 'https://unpod.ai/api/v2/platform';
const API_TOKEN = process.env.UNPOD_API_TOKEN;
const SPACE_TOKEN = process.env.UNPOD_SPACE_TOKEN;
const WEBHOOK_URL = process.env.WEBHOOK_URL; // your receiver

const headers = {
  Authorization: `Token ${API_TOKEN}`,
  'Content-Type': 'application/json'
};

// Track processed tasks to avoid duplicates
const processedTasks = new Set();

async function fetchCompletedTasks(page = 1) {
  const res = await axios.get(
    `${UNPOD_API}/spaces/${SPACE_TOKEN}/tasks/?page=${page}&page_size=50`,
    { headers }
  );
  return res.data.data;
}

async function processNewCompletedTasks() {
  const tasks = await fetchCompletedTasks();

  for (const task of tasks) {
    if (task.status === 'completed' && !processedTasks.has(task.task_id)) {
      processedTasks.add(task.task_id);

      // Forward to your webhook receiver
      await axios.post(WEBHOOK_URL, {
        task_id: task.task_id,
        run_id: task.run_id,
        status: task.status,
        input: task.input,
        output: task.output
      });

      console.log(`Forwarded task ${task.task_id} to webhook`);
    }
  }
}

// Poll every 30 seconds
setInterval(processNewCompletedTasks, 30_000);
processNewCompletedTasks(); // run immediately on start

import asyncio
import httpx
import os
from datetime import datetime

UNPOD_API = "https://unpod.ai/api/v2/platform"
API_TOKEN = os.environ["UNPOD_API_TOKEN"]
SPACE_TOKEN = os.environ["UNPOD_SPACE_TOKEN"]
WEBHOOK_URL = os.environ["WEBHOOK_URL"]

HEADERS = {
    "Authorization": f"Token {API_TOKEN}",
    "Content-Type": "application/json"
}

processed_tasks: set[str] = set()

async def fetch_tasks(client: httpx.AsyncClient, page: int = 1) -> list[dict]:
    url = f"{UNPOD_API}/spaces/{SPACE_TOKEN}/tasks/"
    r = await client.get(url, params={"page": page, "page_size": 50}, headers=HEADERS)
    r.raise_for_status()
    return r.json()["data"]

async def poll_and_forward():
    async with httpx.AsyncClient() as client:
        while True:
            tasks = await fetch_tasks(client)

            for task in tasks:
                task_id = task["task_id"]
                if task["status"] == "completed" and task_id not in processed_tasks:
                    processed_tasks.add(task_id)

                    await client.post(WEBHOOK_URL, json={
                        "task_id": task_id,
                        "run_id": task["run_id"],
                        "status": task["status"],
                        "input": task.get("input", {}),
                        "output": task.get("output", {})
                    })
                    print(f"Forwarded task {task_id}")

            await asyncio.sleep(30)  # poll every 30s

if __name__ == "__main__":
    asyncio.run(poll_and_forward())

Step 3: Integrate with Your Business Systems

Push to HubSpot CRM

After a call, update the contact record with the outcome and schedule follow-up:

const hubspot = require('@hubspot/api-client');
const client = new hubspot.Client({ accessToken: process.env.HUBSPOT_TOKEN });

async function updateHubspotContact(input, profile, summary) {
  // Find contact by phone number
  const searchResult = await client.crm.contacts.searchApi.doSearch({
    filterGroups: [{
      filters: [{
        propertyName: 'phone',
        operator: 'EQ',
        value: input.contact_number
      }]
    }],
    properties: ['firstname', 'phone', 'hs_lead_status']
  });

  if (!searchResult.results.length) return;

  const contactId = searchResult.results[0].id;

  // Update contact with call outcome
  await client.crm.contacts.basicApi.update(contactId, {
    properties: {
      hs_lead_status: mapOutcomeToHubspotStatus(profile.outcome),
      notes_last_contacted: new Date().toISOString(),
      hs_sales_email_last_replied: summary.callback_datetime || '',
    }
  });

  // Create an activity note
  await client.crm.objects.notes.basicApi.create({
    properties: {
      hs_note_body: `AI Call Summary:\n${summary.summary}\n\nTone: ${profile.tone}\nInterest: ${profile.interest_level}\nNext Action: ${profile.next_action}`,
      hs_timestamp: new Date().toISOString()
    },
    associations: [{
      to: { id: contactId },
      types: [{ associationCategory: 'HUBSPOT_DEFINED', associationTypeId: 202 }]
    }]
  });
}

function mapOutcomeToHubspotStatus(outcome) {
  const map = {
    'Meeting Scheduled': 'IN_PROGRESS',
    'Not Interested': 'UNQUALIFIED',
    'Connected': 'OPEN',
    'Not Connected': 'OPEN'
  };
  return map[outcome] || 'OPEN';
}

Send Slack Notification

Alert your sales team when a high-interest call completes:

import httpx
import os

SLACK_WEBHOOK = os.environ["SLACK_WEBHOOK_URL"]

async def notify_slack(input: dict, profile: dict, summary: dict, recording_url: str):
    interest = profile.get("interest_level", "Unknown")
    outcome = profile.get("outcome", "Unknown")
    next_action = profile.get("next_action", "Review call")

    # Only alert for high-interest calls
    if interest not in ("Very Interested", "Interested"):
        return

    emoji = "🔥" if interest == "Very Interested" else "✅"

    message = {
        "blocks": [
            {
                "type": "header",
                "text": {"type": "plain_text", "text": f"{emoji} Hot Lead Call Completed"}
            },
            {
                "type": "section",
                "fields": [
                    {"type": "mrkdwn", "text": f"*Contact:*\n{input.get('name')} ({input.get('contact_number')})"},
                    {"type": "mrkdwn", "text": f"*Company:*\n{input.get('company_name', 'N/A')}"},
                    {"type": "mrkdwn", "text": f"*Interest:*\n{interest}"},
                    {"type": "mrkdwn", "text": f"*Outcome:*\n{outcome}"}
                ]
            },
            {
                "type": "section",
                "text": {"type": "mrkdwn", "text": f"*Summary:*\n{summary.get('summary', 'N/A')}"}
            },
            {
                "type": "section",
                "text": {"type": "mrkdwn", "text": f"*Next Action:* {next_action}"}
            },
            {
                "type": "actions",
                "elements": [
                    {
                        "type": "button",
                        "text": {"type": "plain_text", "text": "Listen to Recording"},
                        "url": recording_url
                    }
                ]
            }
        ]
    }

    async with httpx.AsyncClient() as client:
        await client.post(SLACK_WEBHOOK, json=message)

Write to PostgreSQL

Persist call data for analytics and reporting:

import asyncpg
import json
from datetime import datetime

async def save_call_to_db(pool: asyncpg.Pool, task: dict, output: dict, post_call: dict):
    profile = post_call.get("profile_summary", {})
    summary = post_call.get("summary", {})

    await pool.execute("""
        INSERT INTO call_records (
            task_id, run_id, contact_name, contact_phone,
            company, call_status, duration, outcome,
            interest_level, tone, next_action,
            summary_text, recording_url, transcript, created_at
        ) VALUES ($1,$2,$3,$4,$5,$6,$7,$8,$9,$10,$11,$12,$13,$14,$15)
        ON CONFLICT (task_id) DO NOTHING
    """,
        task["task_id"],
        task["run_id"],
        task["input"].get("name"),
        task["input"].get("contact_number"),
        task["input"].get("company_name"),
        output.get("call_status"),
        output.get("duration"),
        profile.get("outcome"),
        profile.get("interest_level"),
        profile.get("tone"),
        profile.get("next_action"),
        summary.get("summary"),
        output.get("recording_url"),
        json.dumps(output.get("transcript", [])),
        datetime.utcnow()
    )

Step 4: Handle Transcripts

The transcript array gives you the full conversation:

def extract_transcript_text(transcript: list[dict]) -> str:
    lines = []
    for msg in transcript:
        role = "Agent" if msg["role"] == "assistant" else "Customer"
        lines.append(f"{role}: {msg['content']}")
    return "\n".join(lines)

def get_call_duration_seconds(output: dict) -> int:
    return output.get("duration", 0)

def was_call_answered(output: dict) -> bool:
    return output.get("call_status") == "completed" and output.get("duration", 0) > 5

Production Checklist

Before going live with your integration:

Idempotency - Use task_id as a deduplication key. Store processed IDs in Redis or DB to prevent double-processing
Error handling - Retry failed webhook deliveries with exponential backoff
Persist cursor - Store the last processed timestamp or task ID so restarts don’t reprocess all history
Secrets - Store UNPOD_API_TOKEN in environment variables, not in code
Rate limits - Unpod API has rate limits; add delays between polling requests if processing large backlogs
Logging - Log every task ID processed for audit trail
Alerting - Alert if the polling service stops delivering (gap in processed timestamps)

Quick Reference: Key Data Fields

Field	Path	Use Case
Contact name	`input.name`	CRM lookup
Phone number	`input.contact_number`	CRM lookup
Call outcome	`output.post_call_data.profile_summary.outcome`	Lead status
Interest level	`output.post_call_data.profile_summary.interest_level`	Prioritization
Next action	`output.post_call_data.profile_summary.next_action`	Task creation
Call summary	`output.post_call_data.summary.summary`	CRM note
Callback time	`output.post_call_data.summary.callback_datetime`	Schedule follow-up
Recording URL	`output.recording_url`	Audit / QA
Transcript	`output.transcript[]`	Analysis / training
Duration	`output.duration`	Billing / analytics

What’s Next

Create Task API

Trigger outbound calls programmatically.

Get Tasks API

Fetch task results and call data.

Call Logs API

Query historical call logs.

Analytics API

Aggregate usage and performance metrics.

Getting Started

API Lists

Webhook Integration: Receive Unpod Call Data in Your System

What Data Is Available After a Call

Integration Architecture

Step 1: Set Up Your Webhook Receiver

Step 2: Build the Polling Service

Step 3: Integrate with Your Business Systems

Push to HubSpot CRM

Send Slack Notification

Write to PostgreSQL

Step 4: Handle Transcripts

Production Checklist

Quick Reference: Key Data Fields

What’s Next

Create Task API

Get Tasks API

Call Logs API

Analytics API

​What Data Is Available After a Call

​Integration Architecture

​Step 1: Set Up Your Webhook Receiver

​Step 2: Build the Polling Service

​Step 3: Integrate with Your Business Systems

​Push to HubSpot CRM

​Send Slack Notification

​Write to PostgreSQL

​Step 4: Handle Transcripts

​Production Checklist

​Quick Reference: Key Data Fields

​What’s Next

Create Task API

Get Tasks API

Call Logs API

Analytics API

What Data Is Available After a Call

Integration Architecture

Step 1: Set Up Your Webhook Receiver

Step 2: Build the Polling Service

Step 3: Integrate with Your Business Systems

Push to HubSpot CRM

Send Slack Notification

Write to PostgreSQL

Step 4: Handle Transcripts

Production Checklist

Quick Reference: Key Data Fields

What’s Next