Skip to main content
Available now. The messaging layer is a first-party Naive Managed Agents capability. Inbound SMS is verified, stored and wakes the number’s agent.
An identity can own a phone number for SMS. US messaging runs over carrier-registered (A2P) messaging, which requires a registered brand and campaign before outbound texts are allowed. Managed Agents handles the sequence; you choose a registration tier.

Provisioning tiers

Fastest path, no business entity required. Managed Agents verifies your identity with an SMS one-time-passcode, then registers a sole-proprietor brand and campaign.
CLI
If the brand is still verifying, provisioning returns pending_brand_verification; re-run after you complete the OTP.

Inbound vs outbound

Outbound texts before campaign approval are refused with a compliance_pending error. Inbound works right away, so an agent can start receiving before it can reply.
Campaign approval arrives asynchronously and unlocks outbound automatically — no re-provisioning needed.

Sending and receiving

CLI
Inbound texts wake the number’s owning agent, the same way inbound email does — see Inbound events. The woken agent cannot answer by text: no agent tool sends SMS yet.

From an agent

A session whose identity owns a number is offered phone.numbers and phone.read. phone.send is offered when outbound messaging is configured, asks a person by default, and is refused with compliance_pending until the campaign is approved. A deployment with no price for a text refuses the send before it leaves, the same way POST /v1/identities/{id}/phones/{pid}/send does. The SDK is Phones; the CLI is Phone.

Next: connections

Act through third-party OAuth apps.