> ## Documentation Index > Fetch the complete documentation index at: https://www.meilisearch.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Sync data to Meilisearch with Kestra > Keep Meilisearch in sync with your data sources using declarative Kestra workflows. [Kestra](https://kestra.io) is an open-source orchestration platform that runs data pipelines from declarative YAML. The [Kestra Meilisearch plugin](https://kestra.io/plugins/plugin-meilisearch) lets you index documents into Meilisearch as a first-class workflow task, so keeping your search index current becomes an observable, retryable, scheduled job instead of a script you have to babysit. ## Why orchestrate the sync A cron job calling a custom script works right up until it doesn't: a network blip leaves your index half-updated, nobody notices the job died days ago, and there is no record of what synced when. Kestra makes the sync a first-class workflow. Every run is logged, retried on failure, observable in a UI, and triggered by a schedule or an event. You describe what should happen, and Kestra handles execution, state, and failure. ## How the pipelines work Most guides in this series follow the same arc: a one-shot **backfill** for the initial load, then an **incremental sync** that keeps Meilisearch current as data is inserted, updated, and deleted. The Elasticsearch and OpenSearch guides are migrations, pairing the backfill with a safe parallel-run cutover. Three ideas recur throughout: * **ION is the handoff format.** Every Kestra extractor writes [ION](https://amazon-ion.github.io/ion-docs/) to internal storage, and the Meilisearch `DocumentAdd` task reads exactly that. The pipelines are just "extract, convert, index" with no glue code. * **Upserts are idempotent.** `DocumentAdd` is add-or-replace, so overlapping sync windows and accidental re-runs are always safe. * **Deletes need explicit handling.** `DocumentAdd` never removes documents. Each guide covers the delete strategy appropriate to its source (soft deletes plus `documents/delete-batch`, tombstone events, or an index alias swap). Store credentials as Kestra secrets (`{{ secret('NAME') }}`), never as plaintext. Inline values in these guides are for readability only. ## Guides Each guide targets a [Meilisearch Cloud](https://www.meilisearch.com/cloud?utm_campaign=oss\&utm_source=docs\&utm_medium=kestra-integration) project (fully managed, nothing to host). You supply your project URL and Default Admin API key. Relational database. Scheduled lookback window on `updated_at`, soft deletes via delete-batch. Document database. Scheduled lookback on an `updatedAt` ISO string, soft deletes via delete-batch. Object storage. Event-driven trigger on a prefix with exactly-once processing. Event stream. Real-time trigger, one execution per message. HTTP API. A `?updated_since` filter with a KV watermark, or a scheduled full re-index. Search engine migration. Scroll the whole index, then cut over with a parallel run. Search engine migration. Scroll the whole index, then cut over with a parallel run. AMQP queue. Real-time trigger, one execution per message, delete events routed to delete-batch. E-commerce platform. Native `updatedAtMin` incremental sync, webhooks for real-time and deletes.