Multi-database pipelines for individuals and small teams

Turn databases, files, and APIs into datasets you can use.

Connect your sources on a visual canvas. Preview a sample, publish, and run to create a dataset. Enable its read-only Query API when needed.

Try the order example View product demo

Start free with five sample orders. Create, preview, publish, and run your first pipeline.

12 database, file, and API input types
Failures preserve the latest successful result
Dataset and Query API delivery

Order data cleaning

Data flowingIngest · Process · Deliver

Sales details.csvCSV file
Customer databasePostgreSQL
Orders endpointHTTP API

FlyETLIngest · Process · Deliver

Filter
Join
Quality check

Orders datasetSuccessful result ready to reuse
Query APIReady for applications to read

Start with the data you already have

CSVExcelPostgreSQLMySQLSQL ServerOracleClickHouseApache DorisStarRocksMaxComputeSnowflakeHTTP API

Built for teams solving these problems

Make your data work repeatable

Connect the sources you already use and make recurring data work repeatable.

01
BI and data consulting
Turn client databases, files, and APIs into datasets you can preview and rerun.
02
Stores and branches
Combine orders, customers, and operating data from multiple business systems.
03
Early SaaS and AI teams
Prepare data from databases and internal APIs, then let your app read the successful result through a Query API.
Explore data APIs for apps

How it works

01
Connect
Connect databases, files, or HTTP APIs.
02
Shape
Arrange your steps and preview the result.
03
Deliver
Publish, run, and reuse datasets through the Query API.

See the product through an example

5 orders become 3 useful results

Keep completed orders and exclude cancelled ones. After registering, try the same example yourself.

Workflow, run status, data and API response are illustrative

Before and afterFlyETL

Filter: status = completed · Keep Alice, Carol and Eve; exclude Bob and David

Original orders · 5 rowsexample-orders.csv

order_id	customer	status	amount
ORD-1001	Alice	completed	120
ORD-1002	Bob	cancelledExcluded	80
ORD-1003	Carol	completed	240
ORD-1004	David	cancelledExcluded	60
ORD-1005	Eve	completed	160

Completed orders · 3 rowsFiltered results

order_id	customer	status	amount
ORD-1001	Alice	completed	120
ORD-1003	Carol	completed	240
ORD-1005	Eve	completed	160

Start your first preview and run with these orders. No data preparation needed.Try the order example

Current product scale

Connect, process, and deliver at scale

Supported inputs, processing steps, and paid dataset capacity.

12
data input types
Nine databases and warehouses, plus CSV, Excel, and HTTP API.
13
processing capabilities
Including filters, joins, aggregates, deduplication, sorting, quality checks, and custom SQL.
10 million
row-scale datasets
Paid plan limits increase by tier; formal runs process complete input.

Reliability is more than a claim

A failed run preserves your last successful result

Check first, then publish. Only a complete successful run updates the current result, so a failure never turns the last usable dataset into a partial result.

Preview before publishing

Use input samples while editing to inspect fields and processing results. Formal runs still process full data; a sample preview is not full validation.

Failures preserve the last success

If a run fails, is canceled, or times out, the most recent successful dataset content remains available.

Successful results stay deliverable

A complete successful result becomes the current dataset and remains available to business systems and AI applications through the Query API.

Choose by how you work

Prove a real workflow first, then upgrade for more automation

Free supports one lightweight weekly workflow. Paid plans add continuous automation, more run capacity, and team collaboration.

Swipe left or right to compare plans →

Light personal use

Personal Free

Maintain one lightweight weekly workflow and prove its value first.

Free

Start free

Usage: Light weekly automation
Data: 5 datasets · 500 MiB storage · 500,000 rows per dataset
Runs: 15 runs/month · 1 concurrent
Automation: 1 weekly schedule
Delivery: 10,000 queries/month · 7-day history

Best for ongoing use

Ongoing automation

Personal Plus

Keep regular workflows running and trigger them from external systems.

$19/ month

Choose Plus

Usage: Ongoing automation
Data: 100 datasets · 5 GiB storage · 20,000,000 rows per dataset
Runs: 500 runs/month · 4 concurrent
Automation: Minute-level schedules · External triggers
Delivery: 100,000 queries/month · 30-day history

Frequent and heavy workloads

Personal Pro

Run more work at once and keep headroom for heavier datasets.

$49/ month

Choose Pro

Usage: Frequent and heavy work
Data: 200 datasets · 20 GiB storage · 50,000,000 rows per dataset
Runs: 2,000 runs/month · 8 concurrent
Automation: Minute-level schedules · External triggers
Delivery: 1,000,000 queries/month · 90-day history

USD plans renew automatically and can be canceled in subscription settings. Core modeling and publishing stay available on every plan.

Start with a real workflow

Make the next data run repeatable

Create a personal workspace and turn repeated multi-source data work into a clear, reliable, repeatable run.

Try the order example Get help connecting your first dataset

Turn databases, files, and APIs into datasets you can use.

Order data cleaning

Make your data work repeatable

BI and data consulting

Stores and branches

Early SaaS and AI teams

How it works

Connect

Shape

Deliver