TLDR: FreeSWITCH is the open-source engine. SignalWire is the managed cloud built by the same people who wrote it — with a Twilio-compatible API bolted on. Once I understood the lineage, the decision got a lot clearer.

Where I Actually Started

I'm running two telephony stacks right now.

Telnyx powers Apollo's /call skill — a TeXML webhook on Vercel, ElevenLabs TTS piped through a proxy, calls delivered to my cell. It works beautifully. (Telnyx is a carrier and CPaaS platform, like Twilio but with better pricing and a REST-first API.)

Twilio handles a patient outreach tool for a cancer education business — SMS for patient outreach, the 10DLC campaign headaches, the whole thing.

Neither choice was especially deliberate. Twilio was the obvious default. Telnyx came from a recommendation — and honestly, nobody talks about Telnyx even though it's GENUINELY powerful.

That asymmetry is what my telephony consultant called out.

The Call That Reframed Everything

On a 6/19 strategy session with my telephony consultant — he pointed me toward SignalWire.

"The guys that wrote FreeSWITCH wrote SignalWire."

That one sentence reorganized a lot of things in my head.

FreeSWITCH (open-source telephony engine, the thing you'd self-host if you wanted full control over your voice infrastructure) was authored by Anthony Minessale. SignalWire is his managed cloud product, built on that same lineage.

So the tradeoff isn't really FreeSWITCH vs. SignalWire. It's self-host the engine vs. buy the managed platform from the people who built the engine.

That's a very different conversation.

The Part That Got My Attention

His actual pitch wasn't "migrate everything."

It was: look at the SignalWire SDK — they have great examples for wiring a voice agent into an active DOM on a website.

That's the use case I've been circling for months. Real-time voice agents connected to a live web page. Not a phone call. Not a SIP trunk. A voice thread embedded in an app.

And then he mentioned the compatibility layer: SignalWire ships a Twilio-compatible API.

That means I could drop it in next to my existing Twilio code with minimal surgery. It's not a rip-and-replace — it's a dial I can turn.

The Decision I'm Not Making Yet

I'm not migrating anything today.

What I am doing is reading the SignalWire SDK with fresh eyes, specifically for the browser voice agent examples. If the DOM integration is as clean as he says, that opens a whole product path — voice agents embedded directly in the cancer education business's patient outreach interface, not sent through a phone number.

The guy who wrote Asterisk (another foundational open-source telephony engine) is actually a friend of mine. I mentioned that on the call and he didn't blink. This world is small.

Why It Matters to Me

The default in voice infra is to reach for the brand with the best marketing. Twilio has incredible docs and developer love — genuinely earned.

But the signal I keep getting, from people who actually run telephony at scale, is that the interesting platforms are the ones building on the same underlying engines without the brand tax.

Telnyx. SignalWire. Understand the lineage, and the pricing and API surface start making a lot more sense.

That's the thing I'd tell any builder: before you commit to a telephony vendor, ask who actually wrote the engine underneath it.

P.S. To my telephony consultant, if you're reading this — yes, I looked at the SDK. You were right.