Your starter stack got you to scale. Now margin loss, support delays, and compliance friction are slowing you down. Designed for product and engineering leaders, this playbook outlines a phased migration path to Bandwidth’s carrier-owned network to protect your developer roadmap while fixing your call/message economics.
The CPaaS that got you up and running with a credit card at 50K minutes of calls or 500k messages per month is not the same provider you need at 500K minutes or 1M messages. This happens when a stack built out of ease of onboarding starts cracking under the pressure of scale.
Most teams have switching inertia because of very real, competing roadmap priorities. Timing never feels perfect, and there is never reserved development team capacity for a migration in the quarter. It takes them time to realize that they’re in a crunch because they need an enterprise-scale CPaaS switch that’s already overdue, but they cannot afford to derail their current release cycle.
The engineering lift may seem bigger than the savings, but that inertia costs you as you get bigger.
| Myth | Reality |
|---|---|
| Switching means a long-winded rebuild | You can get started with Voice API right away with a migration toolkit. For messaging, If you’re registered directly with TCR as a CSP, you can migrate active 10DLC campaigns with a Connectivity Partner (CNP) without re‑registering or re-vetting. |
| There’s no dev capacity for a migration | A migration toolkit lets you refine your stack at your own pace |
| 60-second billing for voice is just how voice pricing works | On a 1:06 call, you’re paying for 54 seconds of dead air, every call, forever |
| Latency problems are a model problem | A rented network can lower call quality and make good AI sound bad, no matter how well the model works. |
| Switching always requires re-registering messaging campaigns from scratch | While re-registering messaging campaigns is sometimes unavoidable, it’s possible to migrate messaging traffic with their original campaigns intact. |
| Deliverability rates are the same regardless of who you use | Carrier-direct relationships and AI-assisted registration tools change what actually reaches the handset. Not all CPaaS providers have the same path to the carrier. |
When your outbound voice gets billed in full 60-second increments, a call that runs 1 minute and 6 seconds gets rounded up and billed as 2 full minutes. Bandwidth bills in 6-second increments, so that same 1:06 call is billed at 66 seconds, not 120. That gap, at a mid-market call volume, stops being a rounding footnote and starts being a line item.
Latency and support costs strengthen the case for a switch. Voice AI builders don’t just lose customers because a model gave a bad answer. Unnecessary hops in their calls add a beat of dead air before the answer arrives, affecting the Speech-to-Text conversion, and the caller assumes the AI is broken. The intelligence layer inherits whatever the infrastructure gets wrong.
Bandwidth’s direct-to-PSTN connectivity is purpose-built to reduce unnecessary hops.
* Based on Bandwidth network data, North America
Businesses that use messaging as a high-throughput, critical channel of communication turn to Bandwidth for enterprise-scale use and support across RCS as well as different SMS formats: Toll-free, 10DLC, short codes, global 2-way SMS, and alphanumeric SMS. Being a carrier, Bandwidth leverages direct carrier access and upfront AI-assisted registration tools to help reduce typical approval timelines, often to 2 to 3 business days.
Similarly with support. Bandwidth’s Signature Support gives you direct access to the technical guidance and expertise you need via phone, email, or support ticket at no additional cost.
The fear here was never really about switching CPaaS platforms but about tweaking an application that’s actively in use. When you separate those two things, the rest of the transition gets easier.
Bandwidth simplifies the switch for Twilio Voice API users with our Developer Migration Toolkit: an open-source compatibility layer that translates Twilio’s TwiML into Bandwidth’s BXML in real time. This way, your existing Twilio-pattern API calls can test calls on Bandwidth without a rewrite, right up until you’re ready to build natively in BXML.
When it’s time to move, the toolkit is agent native. Just point Claude at the generated code and the Bandwidth CLI. With [band] handling number provisioning, app setup, and callback wiring, your agent completes the full integration, no dashboard required.
For a vertical SaaS team, that’s the difference between planning for a rip-and-replace vs a few minor revisions to otherwise compatible code. For a voice AI builder, it means fewer dev cycles pulled off ongoing product work to deal with under-the-hood infrastructure instead. You can get started right away and plan a phased cutover based on your timeline.
Score your own migration complexity before you talk to anyone. Answer all six questions, add up the points, and check your total against the readiness key at the bottom.
| What’s your current monthly voice call volume? | Your score |
|---|---|
| A. Under 10,000 minutes a month | 1 pt |
| B. 10,000–100,000 minutes a month | 2 pts |
| C. 100,000–1,000,000 minutes a month | 3 pts |
| D. Over 1,000,000 minutes a month | 4 pts |
| How would you describe your current TwiML setup? | Your score |
|---|---|
| A. A handful of simple flows: answer, forward, play a message | 1 pt |
| B. Moderate logic: IVR menus, basic call routing | 2 pts |
| C. 100,000–1,000,000 minutes a month | 3 pts |
| D. Deeply custom: Flex, Segment, or multiple interconnected apps | 4 pts |
| How sensitive is your use case to latency? | Your score |
|---|---|
| A. Not sensitive: asynchronous or notification-only calls | 1 pt |
| B. Somewhat sensitive: standard customer support calls | 2 pts |
| C. Sensitive: real-time conversational AI or voice agents | 3 pts |
| D. Mission-critical: every beat of delay affects the product | 4pts |
| How much engineering time can you realistically dedicate to a migration? | Your score |
|---|---|
| A. A dedicated team for a full sprint or more | 1 pt |
| B. One or two engineers, part-time | 2 pts |
| C. One engineer, squeezed in around other priorities | 3 pts |
| D. None dedicated: this would happen in the cracks | 4 pts |
| Which of these apply to your calls today? | Your score |
|---|---|
| A. None of the below | 1 pt |
| B. Standard TCPA / A2P compliance only | 2 pts |
| C. HIPAA, PCI, or another industry-specific regulation | 3 pts |
| D. 911 / emergency calling obligations, or several overlapping regulations | 4 pts |
| What’s your tolerance for support response time during the switch? | Your score |
|---|---|
| A. High: we can self-serve and troubleshoot | 1 pt |
| B. Moderate: email support with reasonable SLAs is fine | 2 pts |
| C. Low: we need fast, reliable support during the transition | 3 pts |
| D. Zero: any downtime is a business emergency | 4 pts |
| Calculate your score (_/24) |
| Total score | Readiness tier | What it means |
|---|---|---|
| 6–10 | Low complexity | You are set up for a near seamless migration to Bandwidth. Most, if not all your TwiML will convert to BXML with little developer lift required. |
| 11–17 | 2 pts | You’ll need a closer look to determine what’s required for the call flows that don’t have an exact match. Test the call flows in the Real Time translator for a deterministic evaluation that you can actually build on. |
| 18–24 | High complexity | Enough of your stack is involved that a migration engineer should be in the room from day one. That doesn’t mean it isn’t worth doing. It means the plan needs more structure before the first call moves. |
The switch is a sequence you can control, not a rip-and-replace project. The phases below are a guide. The migration completion timeline is dictated by your team’s priorities (and not the other way round). The Migration Toolkit removes the starting barrier.
Start here: Set up a Bandwidth account. If you’re already on the Bandwidth App, you’re good to go. If not, Bandwidth Build is your free developer account. Setup takes minutes; no contract required. You can use your Bandwidth Build account to evaluate Bandwidth’s Voice API on its own, or go ahead and provision a number to test call flows without signing a contract. Start for free →
Compatibility Check: When you’re ready to evaluate a migration to Bandwidth, the Developer Migration Toolkit allows you to test how your current Twilio application would run on Bandwidth. Copy/paste your existing TwiML code into the Compatibility Check, and identify any minor code issues or blockers that could arise in your migration. Then, assess what’s translating cleanly and what may take a bit more work. See Migration Toolkit →
Test your Twilio app with Bandwidth: Test the call flows that you assessed in your Compatibility Check, by using the toolkit’s Real-Time Translator to make your call with Bandwidth. If you hit a feature we don’t support, the call surfaces that loudly. When you’re satisfied, the TwiML to BXML translation you’ve generated as part of your test is captured for the next step in your journey: creating useable BXML code for your migration. See dev docs →
Migrate and scale: Once you’ve tested call flows with the Translator, use the BXML generator to produce a full coverage report, designed for your engineering and ops teams: the BXML works as usable code for your new application.. The BXML generated at this step works as usable code for your new application, and because the Migration Toolkit is agent-native, simply point Claude in its direction to finish the integration with Bandwidth.
If you’ve used Bandwidth Build to get started and you’re ready to scale, you can upgrade to the full Bandwidth App experience when your contract is signed with the same account ID, with no re-integration required. Your voice setup doesn’t change, but your capabilities expand. When it comes to account provisioning (buying numbers, creating voice applications, activation), Bandwidth’s CLI makes it easy to cutover once the dev work is complete. Establish your new support relationship and monitoring baseline before you consider the migration done.
| Network | Owned Tier 1 network with 99.999% core network uptime globally, CLEC in the US, direct carrier relationships | Software layer over third-party carriers | Privately-owned IP network | Owns Inteliquent, fragmented across acquired platforms |
| Billing | 6-second increments | 60-second increments¹ | 60-second increments² | Varies based on product, service agreement, and destination³ |
| Support | Direct access to network engineers | Tiered, chat and ticket-first at scale | Improving, still building an enterprise track record | Chat and email only, no unified support org |
By owning the underlying network instead of renting it, Bandwidth fundamentally changes the economics and reliability of your communications stack.
Talk to a migration engineer and get a personalized cost-of-staying calculation, built on your actual call volume, instead of an industry average.
Or start smaller: Build for free on Bandwidth Voice API and test it out for your use case.
The Developer Migration Toolkit handles voice. Messaging migration runs on a different track, and is more straightforward.
Most of the core logic in your Twilio messaging implementation carries over. Bandwidth’s Messaging API is a builder-first REST API that uses JSON, the same format most developers already prefer. If your Twilio setup uses a REST API (not TwiML) for messaging, you’re already building in the right direction.
The key differences are in callbacks and event handling. You’ll need to update endpoints to accept Bandwidth’s delivery receipts and message events: incoming message, message delivered, and message error. Database tables tracking delivery status will also need to reflect Bandwidth’s event schema.
This is where teams are apprehensive. If your 10DLC campaigns or short code registrations are in good standing with Twilio, it’s possible to migrate that traffic with campaigns intact rather than starting re-registration from scratch. Bandwidth’s carrier-direct relationships and upfront AI vetting tools help accelerate the process when re-registration is unavoidable, often bringing approval timelines down to 2 to 3 business days.
If you register your 10DLC campaigns directly with TCR as Campaign Service Providers (CSPs), you can now migrate approved, active campaigns to Bandwidth as your Connectivity Partner (CNP) without re‑registering or re-vetting. The transition is quick and cost-effective.
Bandwidth covers the full enterprise messaging stack: 10DLC, short codes, toll-free messaging (with handset-level delivery receipts), RCS, global 2-way SMS, and alphanumeric SMS. If your current use case is 10DLC only, there’s also headroom to expand.
Bandwidth’s messaging experts help with code, bringing deliverability strategy and carrier-relationship expertise alongside the API. Reach out to your migration engineer or start with the messaging dev docs.
Deliverability is actually a carrier-relationship question. Carrier experience helps you navigate messaging requirements better and provides deeper insights into what’s really happening to your messages.
Every message travels across an elaborate web of interconnected carrier networks, hitting checkpoints along the way, and the rules at each checkpoint can change at any time. A provider with strong carrier relationships stays ahead of those changes before they impact your traffic. 60% of respondents in Bandwidth’s 2025 State of Messaging Report ranked deliverability as their top factor in choosing a messaging provider. That’s the bar your provider needs to clear.
Day 0 is also the right time to audit your current Twilio messaging setup: document your active campaigns, registered numbers, and messaging formats in use. Voice and messaging run as parallel workstreams from here, each with their own checkpoints. You won’t have to wait until the voice migration is done to start the messaging inventory.
Review your 10DLC campaign registrations and short code or toll-free configurations. Determine which campaigns are ready to move and which may require re-registration.
If you’re your own CSP, you can migrate approved, active campaigns to Bandwidth without re‑registering. So you’re ready to use the full potential of our high-throughput messaging quickly.
Either way, connect with your Bandwidth messaging expert early because campaign portability decisions made in week one can save weeks later.
Update API endpoints to accept Bandwidth’s message delivery callbacks and verify delivery receipts map correctly to your database schema. Campaigns requiring re-registration can typically be approved in 2 to 3 business days with Bandwidth’s automated vetting.
Cut over messaging traffic to Bandwidth. Confirm deliverability baselines against your Twilio benchmarks. Establish monitoring before you consider either workstream done.
Talk to a messaging expert about your 10DLC setup, campaign registration, and which formats fit your use case best.
[1] Twilio, How are Twilio voice and calling product minutes rounded for billing?
[2] Telnyx support, Billing increments
[3] Sinch, Operator Connect Agreement
See how to migrate in minutes
