Migrating from Twilio

Three steps, zero drama

Migrating telephony providers sounds terrifying — phone numbers are identity, and downtime is lost business. But a Twilio migration is a solved problem with a standard shape: inventory what you have, port and adapt numbers and code, then dual-run and cut over when the new side proves itself. esimgomobile ships a guided toolkit for exactly this; this guide is the playbook behind it.

Step 1: know what you have

Before touching anything, catalog every Twilio dependency. Miss one and it pages you at 2 a.m. after cutover.

Numbers. Every phone number: what it's for (voice, SMS, both), where it's published (website, Google Business, print), and whether it's portable. Most Twilio numbers port out cleanly; the exceptions are usually short codes and some toll-free arrangements — identify those early.

Voice flows. Every webhook URL Twilio calls: what TwiML it returns, what logic runs (lookups, branching, recordings), and what happens on failure. Diagram each flow — you'll rebuild them against the new endpoints in Step 2.

Messaging. Every SMS/MMS use: notifications, 2FA codes, marketing, conversations. Note the sending numbers, throughput needs, and any compliance registrations (10DLC, short codes) — these have lead times and don't transfer automatically.

Everything else. API keys and where they're stored, subaccounts, usage triggers and alerts, billing contacts, and the integrations nobody remembers until they break (that one Zapier flow, the CRM plugin). The inventory is also your test plan: every item on it gets a verification check in Step 3.

Step 2a: map concepts, then adapt code

The good news: telephony concepts are universal. A phone number is a phone number; a voice webhook is an HTTP request that returns call-handling instructions; an SMS API sends a message. The migration is translation, not reinvention:

Voice webhooks are the smallest change. Your endpoint already receives call events and returns instructions — you keep the endpoint and adapt the response format to esimgomobile's compatible voice endpoints (final markup publishes at launch). The surrounding logic — database lookups, branching, business rules — doesn't move at all:

# Express voice webhook — before (Twilio TwiML) …
app.post("/voice/incoming", (req, res) => {
  const caller = lookupCaller(req.body.From);   // your logic — unchanged
  res.type("text/xml");
  res.send(`<Response><Say>Hi ${caller.name}, …</Say></Response>`);
});

# … after (esimgomobile compatible endpoint — launching)
app.post("/voice/incoming", (req, res) => {
  const caller = lookupCaller(req.body.from);   // your logic — unchanged
  res.json({ say: `Hi ${caller.name}, …` });   // illustrative shape
});

Messaging follows the same pattern: your send-call sites change the endpoint and payload shape; your inbound-message webhook keeps its URL and adapts parsing. Studio flows are the one case worth rethinking rather than translating — and it's an upgrade opportunity: esimgomobile's Tele Agent handles conversational voice natively, which is what most Studio flows were approximating with drag-and-drop.

Step 2b: move the numbers

Number porting moves your existing numbers from Twilio to esimgomobile with no change for your callers. The process is standardized across carriers:

1. Request. List the numbers to port in the migration toolkit. 2. Authorize. Sign a Letter of Authorization (LOA) per number or range — this is the legal permission for the move. 3. Validate. The losing carrier (Twilio's underlying carrier) confirms the numbers and account details match. Mismatches here are the #1 cause of delays — get the authorized name and address exactly right. 4. Port date. A cutover window is scheduled, usually within days for simple ports. 5. Complete. Calls route to esimgomobile; test immediately.

Two practical rules: don't cancel Twilio until ports complete (canceled numbers can't port), and port in batches — move a low-risk number first, verify end to end, then move the rest. Porting timelines vary by country and number type; the toolkit tracks each port's status so nothing goes quiet.

Step 3: dual-run, verify, cut over

The safe cutover runs both platforms in parallel. New code paths point at esimgomobile while Twilio remains as fallback; test traffic — then a growing share of real traffic — flows through the new side. Your inventory from Step 1 becomes the verification checklist:

Test calls pass end to end on every number: inbound, outbound, SMS both directions, voicemail/message flows. Webhooks verified — signatures validate, retries behave, idempotency holds under duplicate delivery. Monitoring is on: delivery rates, error rates, latency percentiles, with alerts that reach a human. Rollback plan exists and is tested — the old Twilio config restorable in minutes, because the only migration without a rollback plan is a gamble. Team briefed — support knows the new flows, the new dashboard, and who to call.

Then cut over: point DNS/config at esimgomobile, watch the monitors like a hawk for 48 hours, and only then wind Twilio down. Keep the old account in a minimal state for one billing cycle — cheap insurance against the integration you forgot.

Before you ask.