Vapi
BYOC SIP trunk for Vapi assistants. Inbound DID routing, outbound Caller-ID, STIR/SHAKEN A-attestation on US/CA. Sub-50ms regional ingress.
Phone numbers for AI voice agents
DIDHub provisions DIDs for every major AI voice platform — Vapi, Retell AI, ElevenLabs Conversational AI, Bland AI, Synthflow, LiveKit Agents, Pipecat, Vocode, Cartesia, Deepgram Voice Agent. Each integration page below has the step-by-step BYOC setup, Caller-ID configuration, STIR/SHAKEN posture, and latency notes for that platform.
Pick your AI voice platform
Retell AI
Connect a DIDHub SIP trunk to Retell. Custom phone numbers, dynamic agent routing, full call-event webhooks.
ElevenLabs Conversational AI
Phone numbers for ElevenLabs Conversational AI agents. Native BYOC SIP, low-latency interconnect for the realtime models.
Bland AI
Bland AI phone number integration via DIDHub SIP trunk. Outbound caller ID, large-volume ports, STIR/SHAKEN attestation.
Synthflow
Synthflow BYOC: per-DID assignment, inbound to specific Synthflow assistants, outbound from Synthflow campaigns.
LiveKit Agents
SIP trunk for LiveKit Agents framework. Plays nicely with OpenAI Realtime, Cartesia, Deepgram on the LiveKit media plane.
Pipecat (Daily)
SIP-to-Pipecat pipeline integration. Daily.co media transport, low-latency PSTN ingress for the open-source agent framework.
Vocode
Open-source Vocode phone-call agents with DIDHub SIP. Telephony provider config + Caller-ID setup.
Cartesia
Cartesia Sonic + DIDHub for low-latency voice. SIP trunk + media transport notes.
Deepgram Voice Agent
Deepgram Voice Agent API + DIDHub DIDs. STT + LLM + TTS + telephony in one stack.
OpenAI Realtime API
Bridge OpenAI's Realtime API to PSTN via DIDHub + LiveKit / Pipecat / custom SIP-to-WebSocket bridge.
AssemblyAI
Real-time speech-to-text + LeMUR LLM features. SIP-to-WebSocket bridge for PSTN ingress to AssemblyAI's streaming API.
What these AI voice integrations have in common
STIR/SHAKEN A-attestation
US/CA outbound calls from DIDHub-allocated or DIDHub-ported numbers are signed with A-attestation — the only level that reliably avoids "Spam Likely" labels on T-Mobile, Verizon, and AT&T.
Sub-50ms regional ingress
SIP/RTP edges in NOAM (Ashburn, San Jose, Dallas), EU (Frankfurt, Amsterdam), APAC (Singapore, Tokyo), and MENA (Dubai). Pick the closest to your AI inference region.
Bulk DID allocation API
Programmatic provisioning for per-customer DID patterns. POST one DID per onboarding, attach to your AI platform via metadata.
136-country coverage
One BYOC trunk for inbound + outbound across 136 countries. No multi-vendor coordination for international AI deployments.
SIP/TLS + SRTP everywhere
TLS 1.2+ for SIP signaling, SRTP for media. DTLS-SRTP for WebRTC-bridged AI media servers.
Webhook events for everything
Call started, answered, ended, recording-available, transcription-ready — piped to your webhook endpoint with signed payloads.
Don't see your platform?
If your AI voice platform supports BYOC over standard SIP (which most do), DIDHub will work with it. We've integrated with research-stage frameworks and unreleased platforms before. Email [email protected] with the platform name and we'll either confirm it works out of the box or write up the integration.
Ready to get a number?
Pick a DID in 130+ countries from $1.99/month. Activates instantly on most numbers.