ApiTap - A billion rows on a 256 MB worker

A managed extraction engine that runs on a 256 MB worker — and every number is measured, not projected. → 1 billion rows in one endurance run (99.8% success) → 3.7x faster than Airbyte on identical 1.5M-row sync to BigQuery → 12x less memory · 6x less CPU → Configured by clicking, not by writing SQL or YAML Point ApiTap at an API or database — it lands in your warehouse, typed and deduped. 8 sources, 4 warehouses, free dedicated worker. No credit card.

Hey Product Hunt 👋 A few years ago I got hit with a $1,000 Airbyte bill in a single month — just for moving 100GB of data. At $10/GB, every "let me sync this new source" suddenly became a business decision instead of an engineering one. That month is what made me start building ApiTap. What it is: A managed extraction engine that streams APIs and databases to your warehouse. Configured by clicking — no SQL, no YAML, no glue code. What's different: Built with Rust + Apache DataFusion. Every number on our landing is measured, not projected: → 1 BILLION rows in one endurance run (99.8% success) → 3.7x faster than Airbyte on identical 1.5M-row sync to BigQuery → 12x less memory · 6x less CPU → Runs on a single 256 MB dedicated worker What's live today: - Sources: GitHub, Stripe, Salesforce, Jira, Shopify, HubSpot, Postgres, MySQL - Warehouses: Postgres, BigQuery, ClickHouse, Snowflake - S3 coming soon Free dedicated worker to start — no credit card. I'd love your honest feedback on: 1. Is "configured by clicking" believable, or sounds too good? 2. Which source connector should I build next? 3. What's the one thing that would make you NOT trust this for production data? Solo founder, building in public. Will be here all day answering questions — happy to dive into the architecture, the benchmarks, or the journey. Thanks for checking it out 🙏

ApiTap - A billion rows on a 256 MB worker

Replies