About
Built by a data engineer,
for data teams.
rsync.ai exists because building and maintaining data pipelines is still too slow, too expensive, and too engineer-dependent — even in 2026.
Rahul Vishnoi
Creator · rsync.ai · Brussels, Belgium
Senior data engineering manager with a decade of experience building scalable data platforms across logistics, e-commerce, and SaaS. Spent years stitching together Fivetran, Airbyte, custom ETL scripts, and Temporal workflows — and got tired of the same problems repeating: six-week connector backlogs, per-row pricing surprises, and pipelines that only the original author could debug.
rsync.ai is the tool I wished had existed. Describe what you want to move in plain English, review the AI-generated plan, approve the gates, and watch the pipeline run — fully self-hosted, no per-row pricing, source-available under the Elastic License 2.0.
Why rsync.ai exists
The modern data stack has a connector problem. Every new data source means a ticket to engineering, a 6–12 week wait, and a connector that only the author understands. Managed services like Fivetran charge per row at a scale that punishes growth. Open-source alternatives like Airbyte require deep DevOps investment to run reliably.
LLMs changed what's possible. You can now describe a pipeline in plain English, have an AI agent discover the schema, flag PII fields, propose masking rules, and generate a production-ready connector — all before a human has written a single line of YAML.
rsync.ai is that agent, paired with a Temporal-backed runtime, real-time CDC via Debezium, and a self-hosted model so your data never leaves your infrastructure.
Source-available
rsync.ai is source-available under the Elastic License 2.0. You can read the code, self-host it on your own infrastructure, and contribute. The source lives on GitHub.
github.com/rsync-aiGet in touch
Questions, partnership enquiries, or want a live walkthrough? Email hello@rsync.ai or book a 30-minute demo.