Skip to main content
Last updated: July 22, 2026 · API version: v2 · Tested against: openapi.yaml 2.0.0 · 8 min read
The legacy https://api.unpod.ai/api/v1/ host and Authorization: Bearer scheme are deprecated. All current endpoints use the production host https://unpod.ai/, the /api/v2/platform/ path prefix, and Authorization: Token authentication. See the Authentication guide.

Unpod exposes a full REST API so you can build, automate, and integrate AI voice agents into any system - CRMs, helpdesks, internal tools, or custom workflows. This guide walks you through everything from getting your API key to making your first outbound call.

Prerequisites

  • An Unpod account at unpod.ai
  • At least one configured AI agent
  • A telephony bridge with a phone number assigned
  • Your API key (from AI Studio → API Keys)

Step 1: Authenticate

All Unpod API requests use Token authentication. Add your API key to every request header. Many endpoints also require the Org-Handle header (your organization domain handle).
Response:
Keep your key secret - it has full access to your workspace.

Step 2: Get Your Space Token

Spaces are the core organizational unit in Unpod. Most API calls are scoped to a space.
Response:
Save the token value - you’ll use it to scope task and run lookups.

Step 3: List Your Agents

Fetch all agents configured in your organization:
Response:
Note the handle field of the agent (pilot) you want to use for calls.

Step 4: Create a Task (Outbound Call)

A Task triggers an outbound voice call from your AI agent to one or more contacts. Tasks are created inside a space; the request body takes the agent pilot handle and a documents array of contacts.
Response:
The API returns a run_id and task_ids. The agent will call the number within seconds.

Step 5: Track the Run

Once a task is created, Unpod creates a Run - the actual call execution. Poll runs in the space to track status:
Response:
For per-task detail (transcript, recording, outcome) within a run:

Step 6: Fetch Call Logs

After a call completes, retrieve call detail records (CDR) - including transcript, recording, and post-call analysis:
Call logs include: duration, transcript, sentiment, outcome, and any extracted data from your agent’s analysis config.

Passing Context to Your Agent

Use the per-contact context field (inside each documents item) or the task-level context to pass call-specific information. Your agent’s system prompt can reference this data via template variables - useful for personalizing conversations:

Rate Limits


What’s Next

API Reference

Full endpoint reference with request/response schemas.

Authentication Guide

API key management and security best practices.

Agents API

List, inspect, and manage agents programmatically.

Runs & Executions

Track call execution status and outcomes.