Relayion
← Use cases
AI

AI agents that text humans

AI agents are good at reasoning, planning, and calling APIs. What they have lacked is a reliable, two-way channel to reach real people over SMS, one that does not require the human to install anything.

SMS · OUTBOUND
[Agent] Your driver is 8 min away. Reply with gate code if needed.
↩ "Gate 4F"

An AI agent that needs to reach a human over SMS faces a simple problem: it needs a phone number. Not a shared short code, not a virtual number from a pool. A real number with a real SIM, that the recipient will recognise and respond to.

With Relayion, the agent sends an outbound message with the recipient's number and the message body. The paired Android device sends it over the SIM. The recipient replies. The reply arrives at the webhook the agent is listening on. The agent reads it, reasons about it, and responds or escalates or completes the task.

The loop is fully programmable. Delivery events, reply events, and device status events all arrive as webhook payloads. The agent doesn't need to poll. It listens, reacts, and continues.

Relayion supports multi-device routing, which is useful when an agent needs to reach people in different regions through local SIMs. The same API key, different device IDs.

Agent sends a message, waits for reply
// Agent sends
POST /api/v1/outbound
{
"recipientNumber": "+44 7700 900142",
"body": "[Agent] Your order is ready for pickup.
Reply with your ETA."
}
// Inbound reply arrives at your webhook
POST https://your-app.com/webhooks/relay
{
"event": "inbound.received",
"data": { "senderNumber": "+44 7700 900142",
"body": "About 20 minutes" }
}
More use cases

Ready to connect your first device?

Free plan. One device, 500 messages per month. No credit card required.

Get started free →Read the docs