From first sync to finished dashboard.
Cloud-hosted. Live in minutes.
Sign up for rsync.ai Cloud and ship your first pipeline in minutes — plain-English setup, real-time CDC, PII scanning, an AI connector builder for any API, and the models, workflows, charts and dashboards to put the data to work. No per-row pricing, no catalog ceiling.
Build pipelines in plain English
The hardest part of every ETL tool is the setup. rsync.ai replaces config files and DAG editors with a conversation — so the person who understands the data can build the pipeline, not just the engineer who understands the tool.
Anyone on your team can build pipelines
Describe the sync the way you'd explain it to a colleague. The LLM agent turns it into a concrete, reviewable plan.
- No SQL, no YAML, no DAG editors — the agent breaks your request into discrete steps
- Analysts, ops leads, and PMs can ship pipelines without filing an engineering ticket
- Every step is explained back to you in plain language before anything runs
Build the connector you don't have
The source you need isn't in any tool's catalog? Describe it in plain English — rsync.ai reads the API docs and builds a working connector, no engineering ticket.
- Paste a REST or GraphQL docs URL and get a versioned AI-generated connector in minutes, on rsync.ai Cloud
- Auth, schema discovery, and pagination handled — ships as a Docker image, nothing to hand-code
- The same plain-English flow that builds a pipeline builds the connector it needs
Ask rsync.ai
A chat panel inside the product that answers from your connections, schemas, pipelines and runs, or drafts the thing you ask for.
- Ask why a run failed, or describe a sync and get a pipeline to review
- SQL it writes is marked Not run until you open it in the Explorer and run it under your own role
- Reads metadata by default, never your rows — sample rows are off unless your workspace turns them on
Connect your AI tools over MCP
rsync.ai runs an MCP server, so the AI tools your team already uses can work in your workspace.
- Claude Code, Cursor, Gemini CLI or any MCP client can read pipelines, runs, schemas and models
- They can draft pipelines, workflows and dashboards for a person to review and turn on
- Scoped, expiring tokens; metadata only; every call audited
From landed data to a report your team reads
Landing data is half the job. rsync.ai keeps going: query it, shape it into models, chart it, put it on a dashboard, and send it where your team already looks — in the same product that moved it, under the same workspace roles.
A SQL workbench on your connections
Write a query, read the result, then save it, chart it or turn it into a model.
- Previews are capped at 500 rows, with personal columns masked
- Writes need an admin or owner role, and every write is audited
- Export results as CSV, TSV or JSON, or share one to Slack
Tables that keep themselves up to date
A model is a saved query that rebuilds a table on a schedule, or when what it reads has been refreshed.
- Bronze, silver and gold layers that you name and order
- Scheduled SQL passes an approval gate before it runs
- Version history, freshness deadlines and data checks
Reports and alerts on a schedule
Run a query, check a condition, and send the result to the people who need it.
- Start on a cron schedule, after a pipeline finishes, or by hand
- Steps are forms, not code: query, condition, wait, send, run a pipeline, refresh a model
- Send a message, chart or dashboard to Slack or email — and no model is called when it runs
Charts and dashboards, built in
Draw a model or saved query, then put the charts on one live page.
- Bar, line, area, heatmap, scorecard or table; export as PNG, SVG or CSV
- Dashboards with up to four filters; every tile is read live, as the person looking
- Tiles show how fresh their model is, and a dashboard keeps a version history
Governance that happens before data moves
Most pipelines discover a compliance problem after the data has already landed in the wrong place. rsync.ai inverts that: likely personal data is flagged and PII rules apply before data lands, and no pipeline starts until you confirm it.
PII rules set before a single row moves
rsync.ai scans your columns before transfer and flags likely personal data.
- Automatic scan for likely emails, phone numbers, IDs, and addresses
- Per-field rules: mask, hash (SHA-256 or HMAC), drop, or pass-through
- Rules persist across runs, so a column stays protected on every future sync
You approve, the agent executes
Autonomy without control is a liability. The agent stops and asks whenever your request leaves something open, and it never creates or starts a pipeline until you confirm.
- Pipeline creation, table selection and PII rules each wait for your sign-off
- Nothing starts until you say yes. Schema changes are applied and recorded, and a dropped table waits for a person
- Approvals are logged, so you have an audit trail of who allowed what
Credentials encrypted, code you can audit
The credentials that reach your production systems stay locked down — used to run the pipelines you approve and nothing else.
- Source and destination credentials AES-256 encrypted at rest; on a self-hosted install the key is yours
- Never written to logs in plaintext, never shared beyond running your pipelines
- Connector and pipeline code is source-available (rsync.ai Source-Available License), so your security team can audit exactly what runs
Connect to anything — even sources no tool supports
The catalog ceiling is the reason data teams wait months for a connector that may never come. rsync.ai removes the ceiling entirely: if an API has docs, it can become an AI-generated connector, and databases stream changes in real time.
Built-in connectors for common sources
The sources most teams start with work out of the box — and anything without a built-in connector, the Tool Generator builds on demand.
- Databases (Postgres, MySQL, SQL Server, MongoDB), warehouses (BigQuery, Databricks), object storage (S3, GCS, Azure Blob), and Google Sheets
- SaaS sources like Shopify — with Stripe and GitHub in preview, and more added over time
- No built-in connector for your source? Generate one from its API docs — no catalog ceiling
Real-time CDC for Postgres, MySQL, SQL Server, and MongoDB
Log-based change data capture keeps relational destinations current in near real time, not hours behind.
- Debezium-based CDC streams inserts and updates continuously
- SaaS sources sync on a schedule you describe in plain English
- Incremental by default — no full reloads burning compute or budget
Move any file format between cloud storage
Byte-identical copies between S3, GCS, and Azure Blob — no parsing, no re-encoding.
- PDFs, images, Parquet, video, any binary — moved exactly as-is
- SHA-256 integrity check on every transfer
- Metadata catalog tracks name, size, mime type, and last-modified for incremental re-sync
Operate with confidence — fully managed in the cloud
A pipeline you can't see into is a pipeline you can't trust. rsync.ai Cloud gives you full observability and multi-tenant isolation out of the box — and self-hosting in your own VPC is available now, under the rsync.ai Source-Available License.
See exactly what's happening, always
Live insight into every run, plus the traces to debug the one that failed.
- Live row counts, run history, and error alerts
- OpenTelemetry traces with a built-in SigNoz integration
- Every run keeps its history and OpenTelemetry traces
Per-pipeline namespace isolation
Run many pipelines side by side without them stepping on each other.
- Each pipeline writes into its own destination schema (e.g. shopify.orders vs shopify_brand_b.orders)
- Collision detection and ownership gating prevent accidental overwrites
- Per-pipeline schemas and ownership checks keep one team's sync from overwriting another's
Self-hosting in your own VPC
Available now for teams that need to keep everything in-house — one command, full stack, your infrastructure.
- Run it inside your VPC — on a single VM with Docker Compose, or on Kubernetes with Helm (in preview)
- Credentials AES-256 encrypted at rest with a key you control
- Ollama supported for fully local LLM inference — prompts never leave your network
Frequently asked questions
Do I need to write SQL or YAML to build a pipeline with rsync.ai?
Which APIs can the AI connector builder handle?
Does rsync.ai support real-time change data capture?
How does rsync.ai protect sensitive data and PII?
Can I run rsync.ai on my own infrastructure?
Build your first pipeline this week.
One sentence to a running pipeline — everything on this page is live on rsync.ai Cloud today, except what is marked in preview. Self-host it under the rsync.ai Source-Available License.