Migration
Switching from another provider
A lot of teams are re-examining their proxy vendor right now, and most of them are discovering that the hard part is not choosing a replacement — it is moving without breaking a pipeline that currently works. We handle the move ourselves: an engineer maps your setup, we mirror it on our gateway, and both providers run side by side until your own dashboards make the decision.
Map returned in 2 business days · No cancellation required to start
- No cancellation before you start
- Parallel running means your current provider stays in place until you decide otherwise.
- No minimum during evaluation
- The monthly minimum starts when you commit, not during the parallel period.
- Your integration, unchanged
- We mirror endpoint and session syntax so migration is a config change, not a rewrite.
- A named engineer, not a queue
- One person owns your migration end to end and is reachable directly for its duration.
Four steps, and you control the last two
This is a real sequence with real dependencies, so it is numbered. Steps one and two are ours. Steps three and four happen at whatever pace your measurements justify.
- 01
We map what you have
Day 1 – 2Send us your current endpoints, credential format, targeting parameters, and a week of usage data. An engineer reads it and returns a written map of your setup against ours: which pools correspond, which parameters translate directly, which need a change, and what your volume costs at our published rates. No call required to get this.
- 02
We build your gateway to match
Day 2 – 4Where we can absorb the difference, we do. We mirror your existing endpoint structure, session syntax, and authentication format on our gateway so your code does not change. Where a parameter has no equivalent, we say so in writing rather than silently dropping it.
- 03
You run both in parallel
Week 1 – 3Split a percentage of live traffic to us and keep the rest where it is. You compare success rates, latency, and block rates on your own dashboards against a provider you already trust. We do not ask you to cancel anything to start, and the parallel period is not billed against your minimum.
- 04
You shift the rest when the numbers say so
Week 3 – 4Move the remaining traffic at whatever rate your data justifies. Because both paths stay live throughout, there is no cutover window and nothing to roll back. If the numbers do not justify it, we would rather you stayed put than churned in six months.
What an engineer needs from you
Four things, and none of them require access to your systems. Send them in whatever form you already have; we do not ask you to fill in a template.
- Your current gateway endpoints and credential format
- The targeting parameters you use, and how you pass them
- A week of usage data: volume, success rate, and target mix
- Any address allowlists a third party holds on your behalf
What you get back
A written migration map: pool-for-pool correspondence, parameter translation table, anything that does not translate and what to do about it, and your current volume priced at our published rates. It is a document, not a slide deck, and it arrives whether or not you take a call.
Zero downtime, because nothing is ever switched off
There is no cutover window in this process. Both providers stay live, you move traffic by percentage, and every increment is reversible by changing the same number back.
Start at whatever share you are comfortable with — most teams begin at 5% and hold it for a few days. You measure success rate, latency, and block rate on your own instrumentation, against a baseline you already trust, on identical targets. We provide the same metrics from our side so discrepancies surface early rather than becoming an argument later.
Because your existing provider is still serving the remainder, a problem on our side degrades one slice of traffic rather than stopping a pipeline. If we cannot match your baseline, you turn the percentage back down to zero and you have lost a fortnight of engineering attention rather than a quarter of collection.
We will also tell you when the numbers do not support switching. A customer who moves for the wrong reason churns within two renewals, which is worse for us than the deal we did not sign.
Migration questions
- Do I have to cancel my current provider to start?
- No. The parallel period exists precisely so you do not have to. Your current contract stays in place until your own measurements justify moving, and we do not ask for a termination notice as a condition of starting.
- Will I have to change my code?
- Usually not. We mirror your existing endpoint structure, session syntax, and authentication format on our gateway, so the change is a hostname and a credential. Where a parameter has no equivalent on our side, we tell you in writing before you start rather than dropping it silently at runtime.
- How long does a migration take?
- The map comes back in two business days and the gateway is ready within four. The parallel period runs as long as you want it to; most teams reach a decision inside three weeks because that is how long it takes to gather comparable numbers.
- What does the parallel period cost?
- You pay published rates for the traffic you actually send us, and the monthly minimum is waived for the duration. Migration engineering itself is not billed.
Send us your current setup
Endpoints, parameters, and a week of usage data are enough to produce the map. An engineer reads it and replies within two business days.
One business day, from a named engineer.