- Why connector counts mislead
- Managed, code-first or open source
- Open-source and code-first tools you run yourself
- Managed ELT services
- Streaming and change data capture
- One line on each tool
- My advice if you're thinking of running it yourself
- A write-down-your-own-sources checklist
- Frequently asked questions
Choosing a replacement for an open-source data integration platform comes down to one question first: do you want to run the pipelines yourself, or hand them to a managed service? Settle that, then check every source and destination you need by name, because connector counts on vendor pages are counted differently and rarely tell you what you'll get.
Self-run or managed is the first decision when you're moving off Airbyte, which offers both an open-source platform for your own VPC and a cloud service. Running pipelines yourself gives you control over code, hosting and where data travels, but someone on your team owns every upgrade, failed sync and broken connector. A managed service takes that work away, in exchange for living with the vendor's connector list and release schedule.
Here are nine options, in no particular order, grouped by approach: open-source and code-first tools, managed ELT services, and a streaming tool for real-time change data capture. There are no giant brands among them; most are small, focused companies.
Why connector counts mislead
Connector counts are the first number most vendors in this field put on their home page, and they're among the least useful. The platform you're leaving claims 700+ sources on its page; Portable's page claims 1800+ data sources. Both are vendor claims, and they aren't counted the same way. One vendor may count every endpoint of an app, another may count the whole app once, and some lists include community-built connectors with very different levels of upkeep.
A bigger number doesn't tell you whether the one source you need syncs the tables and fields you use, how often it runs, or what happens when that source changes its API. A short list that covers your six sources well beats a long list that covers five of them badly.
So treat counts as a rough signal of focus, nothing more. A long-tail specialist suits teams with lots of niche SaaS sources. A tool built around databases and change capture suits teams moving mostly from production systems.
Managed, code-first or open source
| Tool | Approach | Good for |
|---|---|---|
| Dataddo | No-code data pipelines | Moving ERP, database and SaaS data without code |
| dltHub (dlt) | Python pipeline library | Building pipelines in Python, with a platform to deploy them |
| Etleap | Iceberg pipelines | Ingesting and transforming data for Iceberg pipelines |
| Meltano | Open-source ELT | Vendor-neutral, open-source data movement into a warehouse |
| Portable | Managed ELT | Connecting long-tail sources to a warehouse |
| Rivery | Managed cloud ELT | Fully managed ELT and data pipelines |
| Skyvia | No-code integration | No-code integration, backup and SQL access |
| Sling | Open-source ELT CLI | Scriptable command line ELT |
| Streamkap | Real-time CDC | Streaming database changes into warehouses and apps |
Open-source and code-first tools you run yourself
Sling is an open-source ELT command line tool, so pipelines become scripts you can version and schedule like any other job. dltHub (dlt) is an open-source Python library for building pipelines in code, with a platform to deploy them when you'd rather not wire up the hosting yourself. Meltano is an open-source, vendor-neutral platform for moving data into a warehouse and beyond, built around the idea that the data and the pipelines are yours.
All three reward teams that treat pipelines as software: code review, tests, version control and a release process. They're less forgiving if nobody on the team enjoys debugging a failed run at the end of a long day.
Managed ELT services
Rivery is a fully managed cloud ELT and data pipeline platform. Dataddo moves and manages data across cloud, on-premise and hybrid systems into warehouses and AI systems, with ERPs, databases and SaaS apps as sources. Skyvia combines no-code integration with backup and SQL access, which suits small teams that want one tool for several jobs. Etleap focuses on ingestion, transformation and observability for Iceberg pipelines, a good fit if your lake is built on that table format. The long-tail specialist from the connector section belongs here too.
With any managed service, ask what happens when a source's API changes, who fixes a broken connector and how quickly, and whether you can see logs and row counts for every sync without opening a support ticket.
Streaming and change data capture
Streamkap handles real-time change data capture and stream processing, sending database changes into warehouses and applications as they happen. That's a different job from a nightly batch load: it suits teams whose dashboards, product features or operations depend on data that's minutes old rather than a day old. If batch is fine for you, a streaming tool adds moving parts you don't need.
One line on each tool
Listed A to Z.
Dataddo
Best for: Companies moving data from ERPs, databases and SaaS apps across cloud, on-premise and hybrid systems into warehouses and AI systems, with secure data movement as the vendor's main pitch.
dltHub (dlt)
Best for: Python developers who want an open-source library for writing pipelines in code, plus a platform for deploying them once they're built.
Etleap
Best for: Data teams building Iceberg pipelines who want ingestion, transformation and observability in one platform.
Meltano
Best for: Engineers who'd like open-source, vendor-neutral data movement into a warehouse and beyond, on the idea its page sums up as data you own.
Portable
Best for: Analysts with lots of niche SaaS sources who want a managed ELT tool focused on long-tail connectors feeding a warehouse.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Rivery
Best for: Teams that would rather host nothing and want a fully managed cloud ELT and data pipeline platform.
Skyvia
Best for: Small teams wanting no-code integration, backup and SQL access together in one cloud data platform.
Sling
Best for: Command line fans who want an open-source ELT tool they can script, which its page title describes as free and open source.
Streamkap
Best for: Product and analytics teams that need database changes captured and streamed in real time into warehouses and applications.
My advice if you're thinking of running it yourself
Be honest about who'll be on call. Open-source and code-first tools are a great fit when you've got engineers who like owning pipelines as code and reviewing changes like any other software. If the person running pipelines is an analyst with a dozen other jobs, a managed or no-code service will save you more evenings than any feature.
Whatever you pick, keep transformations in your warehouse or a separate modelling layer rather than locked inside the pipeline tool. That way the next move, if there is one, means swapping the loader rather than rewriting your logic. My guides to ETL tools and Fivetran alternatives cover more managed options if none of these fit.
A write-down-your-own-sources checklist
- List the five or six sources and destinations you can't live without, by name.
- For each source, note the exact tables, objects and fields you sync today, including custom fields.
- Write down how fresh each one needs to be: daily, hourly or within minutes.
- Mark any source that needs change data capture rather than a full reload.
- Decide now whether you'll run open source yourself or want a managed service.
- In every trial, check each item on the list by name, and tick it off only when real data lands.
- Note what happens when a sync fails: who gets told, and how fast.
If a trial can't tick off your own list, its connector count doesn't matter.
Do you make data pipeline or ELT tools? Tool makers can ask to be considered for a featured place on this list.
Want your tool on this list? Featured spots and reviews here are paid partnerships. Here's how to partner with me.
Frequently asked questions
What's the difference between ETL and ELT?
ETL transforms data before it's loaded into the warehouse. ELT loads raw data first and transforms it inside the warehouse, which is how most modern pipeline tools work.
Is open-source data integration free to run?
The software may be free to use, but you'll still need servers, monitoring and someone to keep it running. Count the hours your team will spend, not just the licence.
What's change data capture?
Change data capture reads the changes made in a database, such as new, updated and deleted rows, and streams them onwards, so the destination stays close to real time without full reloads.
Do connector counts matter?
Only loosely. Vendors count connectors in different ways, so check the sources you need by name and test that they sync the tables and fields you use.
Can analysts run data pipelines without engineers?
With a no-code or managed tool, often yes, for common SaaS and database sources. Code-first and open-source tools usually need someone comfortable with Python, the command line or hosting.
How do I move off a data integration platform safely?
Run the new pipelines alongside the old ones for a while, compare row counts and key fields, then switch reports over one source at a time.
