Which data integration tools don't charge per row?
Most pricing pages hide the billing unit inside a paragraph. This page makes the unit explicit, because the unit — not the number — decides whether your bill is predictable.
A tool only "doesn't charge per row" if rows, gigabytes and connector counts are absent from the contract. Bridge bills a plan plus execution capacity (nodes, vCPU, environments) and enforces a 3× monthly cap; rows, GB, connectors and tasks are explicitly not billing dimensions.
Every billing unit, and what it does to predictability
Two platforms can both say "predictable pricing" and still behave completely differently under load. The table maps each common billing unit to what actually moves the number.
| Billing unit | What scales it | At month 12 of a growing pipeline | Predictable? |
|---|---|---|---|
| Rows / MAR | Inserts + updates + deletes in the window | Grows with churn and sync frequency | Low — multiplicative |
| Gigabytes | Volume counted at capture and sometimes delivery | Grows with volume, doubled on some paths | Medium |
| Events | Messages or records ingested | Grows with volume and retries | Medium |
| Connectors | Number of sources or destinations | Grows with every new system | High, but disconnected from value |
| Tasks / jobs | Number of scheduled jobs | Grows with pipeline count | Medium |
| Capacity (Bridge) | Nodes, vCPU and environments you buy | Flat until you add hardware | High — capped at 3× |
Capacity pricing is not "unlimited". It is bounded: every plan states its node count, vCPU budget and environments, and the engine enforces a hard multiple of the monthly plan price as a ceiling. That combination — explicit capacity plus explicit ceiling — is what makes the invoice forecastable rather than merely affordable.
You are a fit for capacity pricing if…
- ✓Your schemas are change-heavy (soft deletes, frequent updates)
- ✓You sync often enough that the same rows are touched repeatedly
- ✓Procurement needs a maximum number, not an average
- ✓You run the pipeline on your own infrastructure anyway
Capacity pricing is the wrong model if…
- ×Your workload is genuinely tiny and bursty — a free or usage tier may cost less
- ×You need infinite elastic scale without provisioning nodes
- ×You cannot host any component yourself
- ×You want a managed service to own uptime end to end
Billing, answered
Does "no per-row charges" mean unlimited data?
No. It means rows and gigabytes are not metered. Each plan still has a node count, a vCPU budget and an environment limit, and the engine enforces a 3× monthly spend cap. Capacity is explicit; the unit is just not a row.
Is capacity pricing always cheaper?
No. If your dataset is small and rarely changes, a usage tier can be cheaper in absolute terms. Capacity pricing wins when change density or sync frequency multiplies a per-row bill, and it always wins on predictability.
What exactly counts as capacity?
Execution nodes, total vCPU across those nodes, and the number of production and non-production environments. Standard is 1 node up to 4 vCPU with one production environment; Business is up to 2 nodes and 16 vCPU with one production and one non-production environment.
Can I see the arithmetic before I buy?
Yes. The TCO calculator takes your monthly changed rows, sync frequency and connector count and returns both a per-row estimate and a capacity estimate with the cap applied.
Run Bridge in your own environment
Download, activate with a license key, and move your first table in minutes.