Frequently asked questions

Synchronoid is a managed platform for building, running and observing data pipelines. You describe sources and transformations in Python and YAML, and the platform handles scheduling, retries, rate limits, storage and change events.

No. It is in private beta, and every plan needs an invite. You can request access. The current version is v0.14.2. Interfaces can still change between releases, and we list every change in the changelog.

Python 3.10 and newer. Install the package with pip install synchronoid, and the CLI is included. See Install the CLI.

Not during the beta. Synchronoid is a hosted service with regions in Frankfurt and Virginia. Enterprise customers can ask about a dedicated region and private networking.

Airflow schedules tasks and leaves data handling to you. Synchronoid also manages storage layers, provenance, rate limits and change events, and its steps are replay-safe by design. A fair comparison is on the Airflow comparison page.

raw keeps untouched payloads, staged holds normalized and typed facts, and published holds merged, current records with a change history. See Storage layers.

By default raw data is kept for 90 days, staged data for one year, and published data until you delete it. You can change this per asset with sy.Retention(days=...). Run history length depends on your plan.

Every value carries its source, fetch time and run id, so you can see why a value looks the way it does. You can read it with ?include=provenance in the REST API. See Provenance.

Yes. Retries and replays run steps more than once, and the result must be the same as running once. Write by key, avoid reading the system clock and send idempotency keys to outside services. See Retries and replays.

The run resumes from the last finished step. Finished steps keep their results, because steps pass references to stored data.

Steps pass references, not data, so there is no practical payload limit between steps. Single records are limited to 4 MB.

A source that keeps failing is paused automatically, and an operator is alerted. Check the source page in the control room for the error code, fix the cause and run syn sources resume NAME.

No, backfills run at low priority. They do share the source request budget, so use reserve in sy.RateLimit to keep part of it for scheduled runs. See Backfills.

Yes, if both use the same source. The request budget belongs to the source and is shared across all pipelines and workers. See Rate limits.

Use the generic http_json connector with configuration, or write a small custom connector with @sy.connector. Test it with syn connector test. See Sources and connectors.

No, delivery is at least once. Use the event id to ignore duplicates. The details for each channel are in Change feeds.

Secrets are encrypted at rest per workspace. You cannot read them back, only replace or delete them. See Secrets.

Not yet. A SOC 2 Type II audit is in progress, and the audit window opened in Q3 2026. We are GDPR-ready and offer a data processing agreement. More on the Security page.

Explorer is free, Team is $490 per month at the beta price, and Enterprise is custom. Steps that wait for rate-limit budget do not count as step-runs. See Pricing.

Check Troubleshooting first, then use the contact form. Include the output of syn doctor --bundle.