Free tool · nine systems · nothing collected
Pick two systems and find out what we have actually implemented against — not a capability claim, an experience check. It runs in your browser and answers a deliberately narrow question, because the wider one needs data we have not published yet.
Pick two systems. The answer tells you what we have actually built against — not a capability claim we cannot support yet.
This checks our experience, not technical feasibility. The tested capability register — endpoints, rate limits, what breaks at scale — is being written and is not published yet, so this tool deliberately answers the narrower question it can answer honestly.
A tool called "integration checker" ought to tell you whether two products can exchange the records you care about. Doing that honestly needs measured behaviour: real endpoints, real rate limits under load, observed quirks, what breaks at volume, and a date on each.
Half of that now exists. The capability register publishes what each of the nine documents about itself — auth, limits, webhooks, verification, retries, data location — fetched on one day and linked to source. The other half, what we have measured ourselves, does not exist in publishable form, and generating it from vendor documentation would produce a confident-looking tool that was wrong in exactly the places that cost money.
So the tool tells you the one thing we can stand behind: whether each side is in the nine. For what each of those nine actually documents, the register is the page you want — and when the measured column of it exists, this tool becomes the real checker rather than being quietly corrected.
Three accounting packages, two job management systems, one payments platform, one CRM and two automation platforms. That is the complete list of what we have put into production.
A longer list would be more impressive and less true. If your system is not here we will say so rather than imply experience we do not have — which is the same rule that governs what we say about tools generally.
It means we know each side’s authentication, its behaviour under load and where it strains, for those two systems specifically. It means nobody is learning on your money, and it means the estimate has a foundation.
It does not mean a clean join automatically exists. That depends on which records you need to move, in which direction, and how often — and on whether the API access you need sits behind a higher pricing tier, which is the trap on the buyer’s checklist.
Not "can these talk" but "what exactly needs to move, and which way". Most integration disappointments come from a sync that technically works and does not carry the field somebody actually needed.
Three specifics cover most of it: do customers sync both ways or only outbound, do part-payments flow back, and do credit notes flow at all. Those three are also the questions to ask any vendor before you buy, and they are on the accounting page for the same reason.
Bring it to the Leak Check anyway. Plenty of unfashionable products have perfectly good APIs, and plenty of fashionable ones do not. What we will not do is tell you it is straightforward before looking.
If it turns out there is no API and no sensible database access, we will say so and that will be the end of it — which happens, and is better learned in twenty free minutes than three weeks into a build.
There is a middle case worth knowing about: a product with no public API but a supported export, or a database you own and can read directly. Those are workable and we have done them, but they are meaningfully more fragile than a documented API — an export format can change without notice and nothing tells you until the numbers stop matching. We will quote that honestly as the ugly option rather than presenting it as equivalent.
The worst case is a product that is actively hostile to getting your own data out. It exists, it is more common than it should be, and it is the strongest argument on the ownership page — because whatever we build for you, you should be able to leave with it.
No, and it says so on the page. It answers whether we have built against each side, which is a smaller claim we can actually support. For what each system documents about its own limits, webhooks and verification, the capability register has all thirteen fields per system with the gaps marked.
Nine is how many we have implemented against. Every other checker in this category lists hundreds, which should tell you what those lists are made of.
20 minutes. We find where the money's going. No pitch.
Book a free Leak Check