Why BYOC matters
Your carrier pricing, DID inventory, geographic coverage, and regulatory compliance are already solved. Migrating that to a platform-managed carrier means re-porting numbers, renegotiating rates, and accepting reduced flexibility.
Resource — Carriers
Keep your carrier relationship. Connect your AI agent.
Many AI voice platforms want to own your carrier relationship — locking you into their number inventory, pricing, and call routing. BYOC (bring your own carrier) lets you keep your existing SIP carrier and DID inventory while still connecting to any AI voice runtime.
What BYOC requires technically
Your carrier needs to forward inbound calls to a SIP endpoint that can bridge to the AI runtime. That endpoint must terminate SIP signaling, handle RTP media, and adapt the audio format for the agent platform’s WebSocket interface.
How Telepath enables it
Telepath accepts SIP trunks from any standard carrier and bridges inbound calls to your AI voice runtime over WebSocket. Your carrier sends SIP to Telepath. Telepath handles the rest. No number porting, no agent lock-in.
The real reasons to keep your own carrier
Teams that investigate BYOC are usually doing so for one of four concrete reasons. Understanding which one applies to your situation helps clarify what the architecture actually needs to solve.
Per-minute pricing and volume economics
AI voice platforms that provision carrier numbers typically mark up per-minute rates above what direct carrier agreements offer. If you’re handling thousands of minutes per day, the difference between $0.014/min (platform-managed carrier) and $0.004/min (direct Telnyx or Bandwidth agreement) is material. Beyond rate, direct carrier agreements often include volume commitments, dedicated capacity guarantees, and SLA terms that platform carrier arrangements don’t offer. For high-volume deployments, the business case for BYOC is often made in the first month.
DID inventory and number ownership
If your customers call specific numbers that belong to your business — whether published on your website, printed in materials, or embedded in existing workflows — porting those numbers to a platform-managed carrier is a significant operational event. Numbers can take days to port, and the process carries risk. BYOC allows you to keep your DID inventory exactly where it is. The only configuration change at the carrier is a SIP trunk pointing to Telepath rather than (or in addition to) your existing termination endpoint.
Geographic coverage and regulatory compliance
Carrier coverage for toll-free, local, and international numbers varies significantly. Your existing carrier may have negotiated rates or coverage in markets that a platform’s managed carrier doesn’t match. Regulatory requirements in certain markets (local presence, interconnect regulations, emergency services compliance) may require specific carrier relationships. BYOC preserves the compliance posture you’ve already established.
Multi-carrier failover and redundancy
Production voice infrastructure at scale typically uses more than one carrier for redundancy. If your primary carrier has an outage, you need a secondary path. BYOC architecture supports multi-carrier configurations where you control the failover logic. With Telepath as the bridge, you can point multiple carrier SIP trunks at the same Telepath connection and implement carrier failover at the SIP routing layer, without changing anything in your AI agent stack.
Avoiding agent platform lock-in
The AI voice agent market is evolving rapidly. Platforms that exist today may be acquired, change pricing, or be superseded by better options within 12–24 months. If your carrier is locked to your agent platform, switching agents means also migrating your number inventory and carrier relationship. BYOC with Telepath decouples the carrier layer from the agent layer, so switching agent platforms requires changing only your Telepath WebSocket configuration — not re-porting numbers or re-establishing carrier relationships.
Any SIP carrier
Twilio, Telnyx, Bandwidth, Vonage, Plivo, SignalWire, Flowroute, or your own SIP infrastructure. If it sends standard SIP/RTP, Telepath accepts it.
Any AI agent runtime
Vapi, Bland, Retell, LiveKit Agents, or a custom WebSocket endpoint. Telepath delivers PCM16 audio over WebSocket. Any runtime that accepts streaming audio works.
Diagnostics across both
Per-call attribution covers carrier-side SIP timing, gateway performance, and agent-side TTFS. When something goes wrong, you know which layer to escalate to.
Bring your carrier. Connect your agent. Own the call path.
Telepath is built for teams who want to control their carrier relationship and their AI agent stack independently. Share your carrier and expected volume; we’ll scope the connection path.