How do API connections work?
Updated 2 weeks ago
With an API connection, we call the provider's API on a schedule using the credentials your customer authorized, and normalize the data into our unified models. You then read and write that customer's data through the Bindbee API with the connector's connector_token.
How a connection is set up
Your customer authorizes access through a Magic Link or the Embedded SDK.
Depending on the integration, they sign in with OAuth or provide an API key, token or certificate. Some integrations offer both, and the choice doesn't change the data.
Some integrations are report-based. With Workday RaaS, for example, we read a custom report your customer sets up in Workday.
How data stays up to date
A Production connector syncs every 24 hours. The schedule is set per connector. Contact us to change the frequency or start time.
A Development connector syncs once when it's authorized. After that, you trigger each sync yourself.
You can force a resync at any time, and subscribe to webhooks to hear when a sync finishes.
What an API connection can and can't do
Coverage is whatever the provider's API exposes. When the API leaves out data you need, an SFTP connection can fill the gap. ADP Workforce Now, for example, doesn't offer benefits enrollment through its API, but it's available through its SFTP report.
Writes are often supported through API connections. SFTP connections are generally read-only.
We always pull data from the provider's API. We don't connect through a provider's outbound webhooks in place of its API.
When credentials stop working
Tokens expire on their own, and keys stop working when someone rotates them. The connector then moves to Relink Needed and stops syncing until you relink it.