In July 2020, when we understood what Airbyte would become, we saw that current solutions had a few important issues: lack of customizability and agency for users to make changes themselves, inability to cover the long tail of connectors, and prohibitively expensive pricing, especially for large volume sources like databases.
In October 2021, we addressed this third problem and introduced a vision for what pricing would be: consumption-based with credits where users consume credits through compute hours. Our goal was to separate how databases and APIs were getting charged as databases usually replicate 10x faster than APIs.
In April 2022, we removed the waitlist for Airbyte Cloud. It’s now been three months, and we learned a ton — which brings us to now. Today, we’re announcing volume-based consumption. Credits will now be consumed on millions of rows for APIs and GBs for databases and files instead of compute time. Before going into details for this pricing, let us tell you what we learned and why we picked it.
There is too much variance in throughput across APIs and databases across different syncs. It makes compute-based pricing unpredictable in terms of predicting costs for our users.
This unpredictability is not a good experience for our customers and doesn’t necessarily align to value as they could be charged wildly different amounts for the same throughput depending on the sync.
Compute also makes our pricing hard to understand, although we are trying to be as transparent as possible as a company. Our cost estimator helps a bit in translating, but that’s not enough.
Finally, as a company, we weren't incentivized to improve our infrastructure as it could affect our revenue.
Here were our must-have criteria:
Volume is a common language in this market, and most of our users speak in GBs for databases and files, and in millions of rows for APIs. This is exactly how we will model pricing.
Here is what a single credit ($2.50) will provide:
We won’t charge for normalization.
How does that scale with volume?
We offer discounts when you buy bulk credit packages upfront. The larger the package, the higher the discount. Our cost estimator will give you some insight on how volume discounts make our pricing scale. Talk to our sales team to access these discounts.
Separating pricing between APIs, databases and files enables users to pay for value received depending on the source.
Airbyte is also still far more affordable than the common monthly-active-rows (MAR) model:
See how our pricing works for your specific use case with our latest cost estimator.
Although you might know the size of your databases or the number of rows across your APIs, it’s still very hard for most companies to evaluate the volume of their incremental updates.
We are solving this by making our Data Architect team available to you at any point. The goal is to help you connect all your sources and destinations, do it all for free during a 14-day unlimited credit trial, and then we can give you a cost estimate and the corresponding discount offer by the end of the trial. We’ll also help you predict your incremental load.
We consider any change of pricing to be a critical moment in a company’s evolution. Airbyte has always been a transparent company and we hope that shedding light on everything that went into that pricing change will make us worth your trust.
If you have any questions or any comments on this pricing change, please reach out to us in our chat, Slack or via email at firstname.lastname@example.org
Get all your ELT data pipelines running in minutes with Airbyte.