Skip to content
adapters.io

How does data integration work? Plug in, map, sync

Data integration works in three steps: plug into your source and destination, map fields between the two schemas, then sync records on a schedule or webhook. Adapters, a data integration platform, does all three visually, so your first sync runs in minutes, not sprints.

Try the live demo

3 steps · minutes to first sync · 99.98% sync success

Three steps between "we should sync this" and synced

Every integration on Adapters follows the same shape. Connect two systems, wire their fields together, and let the sync run. No glue code at any step.

STEP 01

Plug in

Pick a source and a destination from 200+ connector pairs. Credentials are scoped and encrypted; read-only where the sync allows it. About a minute per side.

STEP 02

Map

Both schemas appear as port lists. Matching fields auto-plug the moment the pair loads; you rewire any cable by hand and attach transforms with a live preview.

STEP 03

Sync

Test on sample records, then go live. Schedules run from hourly down to every minute, or a webhook fires the sync within seconds of the source changing.

The full walkthrough, first login to live sync

Here is exactly what setting up an integration looks like. Budget ten minutes for your first adapter; most people are faster on the second.

  1. 01

    Pick your pair

    Choose a source and destination, say Stripe and QuickBooks, or Postgres and Snowflake. Authorize each side with OAuth or an API key. Adapters stores credentials encrypted with AES-256 and requests the narrowest scopes the sync needs.

  2. 02

    Review the auto-mapping

    Both schemas load as port lists and matching fields plug themselves together: customer.email lands on Customer.Email. Drag any cable to rewire it. This is the same data mapping tool on every plan, Starter through Enterprise.

  3. 03

    Add transforms where the systems disagree

    Click a cable to attach a transform: convert cents to dollars, reformat a Unix timestamp for your books, split full_name into first and last. A live JSON preview shows the record before and after, so nothing is a surprise in production.

  4. 04

    Run a test sync on sample records

    One click pushes a handful of real records through the mapping and shows a per-record result. You can watch this today in the live demo with canned Stripe data, no account needed.

  5. 05

    Go live on a schedule or webhook

    Pick a trigger: hourly on Starter, every 5 minutes on Growth, every minute on Scale, or webhook-driven where the source supports it. From that point the adapter runs itself, with retries, alerts, and logs watching every record.

What happens when a sync fails

APIs time out, rate limits trip, someone renames a field. A data integration is only as good as its failure handling, so here is ours, in order.

Retries, automatically

Failed records retry with exponential backoff, up to 6 attempts over 24 hours. Idempotency keys mean a retry never creates a duplicate. Rate-limited batches wait for the reset window instead of dropping.

Alerts, immediately

If retries are still failing after 15 minutes you get an email, or a Slack message on Scale, naming the adapter, the record, and the API error. No silent failures, no finding out from a customer.

Logs, per record

Every record keeps its own trace: the payload in, each transform applied, the response from the destination. When something looks off, the answer is one search away, and replaying a fixed record is one click.

Webhooks or polling: how a sync knows to run

A sync starts one of two ways: the source pushes a webhook the moment data changes, or Adapters polls on a schedule. We use webhooks wherever the source offers them and fall back to polling everywhere else. The longer version is in our webhooks vs polling guide.

Comparison of webhook and polling triggers
Trigger Webhook Polling
Latency Seconds after the change Your interval: 1 min to hourly
Source support Only where the API offers events Works with any API or database
On Adapters Automatic where available The fallback, with change detection

See the three steps with your own data

Plug in a pair, map the fields, run a sync. Most teams are live in minutes, from $49 a month.

Try the live demo