Skip to content
RelationsAI
Getting started

RelationsAI developer documentation

Preview

The RelationsAI API lets your software do what your team does in the RelationsAI console: place outbound calls with an AI voice agent, run calling campaigns, keep contacts and do-not-call lists in sync, and receive events when calls end. It is a JSON REST API over HTTPS, versioned under /v1, with test keys for a free, simulated sandbox and live keys for real calls.

5 minutesQuickstartCreate a test key and place your first simulated call.ReferenceAPI referenceEvery operation, parameter, response and error.GuideReceive webhooksGet notified when calls end and campaigns change state.FundamentalsError codesWhat each error means, whether to retry and how to fix it.

What you can build

  • Call from your CRM. Place a call the moment a lead arrives, passing the lead's name and details as call variables, and write the outcome back when the call settles.
  • Run campaigns from your data. Create a campaign, add recipients from your warehouse, start it on a schedule, and pause, resume or stop it from your own tools.
  • Keep data consistent. Sync contacts, consent and the do-not-call list between RelationsAI and your systems, so every call respects the same rules.
  • React to events. Receive signed webhooks when a call ends, when its transcript, recording and analysis are final, and when a campaign changes state.

How it fits together

You build, test and publish agents in the RelationsAI console; the API reads them and uses them to place calls. Every API key belongs to one workspace and one mode: a test key works against a simulated sandbox that never dials anyone or charges anything, and a live key places real calls charged to your workspace. Read Platform overview for the full picture, and API availability for what the API covers today.

Start here

  1. Quickstart: a complete, free test-mode call in about five minutes.
  2. Authentication and API keys: how keys, scopes and modes work.
  3. Test mode: the simulated outcomes you can trigger.
  4. Error handling and Idempotency: how to handle failures and retry without repeating an action, before you go live.
  5. Going live when you are ready for real calls.

How the docs are organised

Getting started explains keys, environments and the core objects. Guides walk through complete tasks. API fundamentals covers the conventions every request follows: errors, idempotency, pagination, rate limits and data formats. The API reference is generated from the OpenAPI contract, which you can also download as JSON or YAML. There is no official SDK yet: every example uses plain HTTP in cURL, JavaScript (Node.js 20+ fetch) and Python (requests).

Docs for AI tools

These docs are also published as llms.txt, llms-full.txt and as Markdown: add .md to any page URL.

Getting help

Email support@relationsai.com and include the request_id of the request you are asking about. See Support for what to include and what never to send.

Platform overview