Integrations for plain-English data pipelines
rsync.ai is a plain-English data pipeline: describe a sync and an LLM agent builds it — you approve each step — with built-in MCP connectors for Postgres, MySQL, SQL Server & MongoDB real-time CDC, the BigQuery & Databricks warehouses, S3, Google Cloud Storage, Azure Blob, and Google Sheets, plus Shopify for SaaS, and an AI Tool Generator for any REST/GraphQL API.
Live MCP connectors with step-by-step setup — PostgreSQL, MySQL, SQL Server, and MongoDB with real-time CDC; the BigQuery and Databricks warehouses; AWS S3, Google Cloud Storage, Azure Blob, and Google Sheets for storage; and Shopify for SaaS, with Stripe and GitHub in preview. Don't see yours? The AI MCP Connector Builder generates a working MCP connector from any REST or GraphQL docs URL in minutes.
Popular pipelines
Step-by-step guides for the most common source-to-destination pairs — auth setup, exact resources synced, and the destination schema rsync.ai produces.
SQL Server → warehouse
View guide →MongoDB → warehouse
View guide →Load into BigQuery
View guide →Load into Databricks
View guide →Shopify → MySQL
View guide →Shopify → PostgreSQL
View guide →Shopify → Google Sheets
View guide →Stripe → PostgreSQL
View guide →Postgres → MySQL
View guide →Postgres → AWS S3
View guide →Postgres → Google Sheets
View guide →MySQL → PostgreSQL
View guide →MySQL → AWS S3
View guide →MySQL → Google Sheets
View guide →AWS S3 pipelines
View guide →Google Cloud Storage
View guide →Azure Blob Storage
View guide →All Shopify integrations
View guide →PostgreSQL → warehouse (CDC)
View guide →MySQL → warehouse (CDC)
View guide →Featured source MCP connector
Each source-and-destination pair has its own page with auth setup, the exact resources synced, sync-mode options (batch or CDC), and the destination schema layout rsync.ai produces.
Sync Shopify orders, customers, products, inventory and fulfillments to any destination — full backfill, incremental updates, and PII safe by default.
Sync Stripe charges, customers, subscriptions, invoices, and payment intents to your database — scheduled batch with cursor pagination, no per-row fees. The Stripe connector is in preview.
Real-time CDC from PostgreSQL via logical replication. Capture every INSERT, UPDATE, DELETE sub-second. Works with RDS, Aurora, Supabase, Neon, and self-hosted Postgres.
Real-time CDC from MySQL via binlog ROW replication. GTID support. Works with MySQL 5.7+, RDS, Aurora, Cloud SQL, and PlanetScale.
All available connectors
MCP connectors are grouped by what they do. The same MCP connector can be a source, a destination, or both, depending on the pipeline you describe.
Databases with real-time CDC
Log-based CDC for transactional sources — logical replication, binlog, and MongoDB change streams. Subsecond change capture, transactional consistency, and replayable event history.
Cloud warehouses & lakehouse
Managed analytics destinations. rsync.ai lands partitioned, typed tables with incremental merge and waits for you to approve the schema before anything is created. BigQuery and Databricks are live in production; Snowflake and Redshift are in preview.
Storage & spreadsheets
Object storage and shareable spreadsheets. Useful as both source and destination for backups, analytics handoff, and ad-hoc reporting. Blob passthrough moves any file byte-for-byte between clouds.
SaaS & API connectors
Shopify is live in production; Stripe and GitHub are in preview. Any other SaaS or API source, the AI Tool Generator builds on demand from its REST or GraphQL docs — auth, schema discovery, and pagination handled.
How a rsync.ai integration runs
Every MCP connector — built-in or AI-generated — follows the same three steps. The chat agent does the heavy lifting; you stay in control of the decisions that matter.
Describe the sync in plain English
Say "sync Shopify orders to Postgres every hour" in the chat. No YAML, no DAG, no Python operator to write.
Approve each gate
rsync.ai walks four approval gates — connection test, schema discovery, PII scan, then final approval. It introspects the source, proposes destination tables with column types and row-count estimates, and waits for your sign-off before anything runs.
Run, monitor, and replay
Pipelines emit OpenTelemetry traces to SigNoz. Every run is replayable from the Kafka + Temporal event log, so a bad sync is a one-click rewind.
Don't see your MCP connector?
Point the AI Tool Generator at any REST or GraphQL docs URL and it produces a versioned MCP connector — auth, schema discovery, cursor pagination, and a Dockerfile included. Most APIs are ready to sync in under five minutes.
More integrations live on the homepage — browse all MCP connectors →