Teller is shutting down their API. Here's how to migrate without losing users.

Teller is withdrawing its API product this week. If you build on it, you need a new connection layer fast, and you need one that won't put you right back here again. Here's what to do.

What transfers and what doesn't

Access tokens are bound to the provider that issued them. A token is a credential scoped to one aggregator's connection to one bank, so nothing you hold from Teller carries to a new provider. Every migration path requires your users to re-link their accounts through a new connection.

You don't have to force a mass re-link

Re-linking is unavoidable. A single forced re-auth event on cutover day is not.

Many teams read "users must re-link" and picture emailing their whole user base to reconnect inside a deadline window. That produces a support spike and drops the users who ignore the email. You can run it differently. Prompt each user to reconnect at their next natural touchpoint instead of all at once.

Sequence re-connects against activity you already see:

• When a user next logs in, show a reconnect prompt before the screen that needs live account data.
• When a stale connection fails a refresh, that failure is the trigger to route the user into the new link flow.

Active users move first because they show up first. Dormant accounts reconnect when they return, rather than counting as churn on a date you picked. Your busiest users stay current, and you spread the re-auth load across weeks instead of one fragile morning.

The caveat: any feature that needs a live connection for a user breaks until that user re-links. Balance checks, transaction refreshes, and payment verification all go down for a given account until its owner reconnects. Plan a fallback, whether a cached last-known value or a clear in-app message.

Migrate up a layer, not sideways

You're already paying the migration cost this week. That makes now the cheapest moment you'll ever have to fix the underlying dependency.

Re-linking users and rewiring your connection layer is the expensive part, and you do it either way. Do that rewire once to route through multiple aggregators, and no single provider's exit takes you down again. Swapping one pipe for another just resets the clock on the same risk.

Crew proved this under a real deadline. Weeks before their beta launch, their existing provider couldn't reach several banks their early users needed. Crew integrated Quiltt in under two weeks, connected to thousands of additional institutions, and launched on time. Routing across multiple aggregators through one API means a single provider going away is no longer your problem to solve from scratch.

What to do this week

Talk to us. You can start small, without a big commitment. Our Builder plan runs $100/mo with no extended contract and no setup fees. Pick your connection layer, then re-link users at their next natural touchpoint.

‍